In her concurring judgment in 2747-3174 Québec Inc. v. Quebéc (Régie des permis d'alcool) Madame Justice L'Heureux-Dubé complains that the Supreme Court of Canada currently lacks a coherent and consistent methodology of legal interpretation. Developing such a methodology, she suggests, is a judicial responsibility, for statutory interpretation is one of several areas in which superior courts have an inherent jurisdiction to fix norms.(1)
Anyone who has read the recent case law of the Supreme Court of Canada dealing with statutory interpretation must certainly agree that the pronouncements of the Court on this subject are confusing and contradictory. However, the problem in my view is not methodological. In fact, the interpretive practice of the Court is sound and is often exemplary; and overall this practice is consistent. While it is true that in particular cases judges emphasize sometimes textual meaning, sometimes intended meaning or purpose, and sometimes compelling policy concerns, in doing so their approach is consistently pragmatic. Using a pragmatic approach, each judge takes advantage of the full range of interpretive resources available to interpreters and deploys those resources appropriately given the particularities of the case.
The problem as I see it is not methodological but rhetorical. Faced with a problem in interpretation, a court must not only draw on its skill and integrity to produce appropriate outcomes; it must also offer coherent and acceptable explanations of how its outcome was reached and these explanations must be grounded in a coherent and acceptable theory of the judicial mandate in interpreting legislation. As L'Heureux-Dubé J. suggests, sketching out such a theory is a job for the courts.
The first purpose of this article is to examine the recent case law of the Court dealing with interpretive issues, focusing in particular on the Court`s own explanations and theories about its interpretive practice. I will try to assess the success of its explanations and the coherence and persuasiveness of its theories. In carrying out this assessment I will do my best to discredit the plain meaning rule and the theory of textualism in which it is grounded. I will also explore the reluctance of courts to abandon the theory of intentionalism and explain why I think they should do so. I will suggest that making legislative intent the sole touchstone of interpretation forces judges to work with a formal and rather empty sense of intention -- one that allows for grand talk about unelected judges deferring to the will of legislature but relies on judicial fiat to determine that will.
I will end the article by endorsing a theory of interpretation that emphasizes the importance of the legislative text, and the importance of legislative intention in so far as this can be known, but places final responsibility for outcomes on the shoulders of interpreting judges. On this theory, the judicial task in interpretation is properly characterized not as giving effect to the meaning of the text or the intention of the legislature, but as solving the interpretive problem facing the court in an appropriate and acceptable way. What makes an outcome appropriate and acceptable is not easy to capture in a formula, because there are many variables involved. However, the fact that norms like appropriate and acceptable are complex and multifaceted and their interaction is subtle does not destroy their character as norms. I will argue that these norms have the power to justify outcomes and therefore the power to guide and constrain judicial reasoning and outcomes.
Part 1 of this article surveys three theoretical strands that appear in recent judgments of the Supreme Court of Canada dealing with statutory interpretation, namely textualism (also known as literalism), intentionalism, and pragmatism.(2) My purpose in this part is to explain and document the competing perspectives of different members of the Court.
Part 2 takes a close and lengthy look at the plain meaning rule,(3) which is the version of textualism preferred by members of the Supreme Court of Canada.(4) In this Part, I set out what I believe are serious difficulties with the plain meaning rule. I do not try to survey or to summarize the extensive scholarly literature exploring the recent revival of textualism in American and Canadian courts.(5) My purpose rather is to focus on three aspects of the rule that particularly concern me. The first is the faulty assumptions about language and reading on which textualism is based. My criticisms of these assumptions, it should be noted, are drawn not from philosophical approaches to language such as deconstructionism or hermeneutics, but are grounded in psycholinguistic studies carried out by linguists and cognitive psychologists over the past twenty-five years.(6) The second aspect of the rule that concerns me is that it encourages interpreters to ask only one question in statutory interpretation cases, namely, what is the meaning of this text. As a result of focussing so narrowly on textual meaning, other interpretive issues are overlooked. My final concern with the plain meaning rule is the arbitrary way it is invoked when dealing with some kinds of legislation, primarily tax and penal legislation, but ignored in other contexts.
Part 3 of the article analyses the modern principle formulated by Elmer Driedger in the 2nd edition of his book Construction of Statutes:
Today there is only one principle or approach, namely, the words of an Act are to be read in their entire context and in their grammatical and ordinary sense harmoniously with the scheme of the Act, the object of the Act, and the intention of Parliament.(7)
This formula has been cited and relied on by Canadian courts on countless occasions. I focus on some overlooked elements of the formula, pointing out what Driedger meant by "entire context" and "intention of Parliament". I also draw attention to his emphasis on the concept of harmony between the words of the text and the general body of law. I try to show that Driedger's modern principle expresses an intentionalist approach to interpretation and that its association with the plain meaning rule in recent case law is the result of confusion. I also suggest that the re-formulation of this principle in the third edition of Driedger is not necessarily a break with intentionalism.
Part 4 of the article offers a pragmatic justification of the actual practice of most Canadian appeal court judges, including the literalists. The pragmatic way of understanding judicial reasoning does not seem to me to be particularly innovative or iconoclastic. It merely draws out the implications of certain unavoidable realities - that legislation is not self-applying, and that determining the meaning of the legislative text is only one of several things that courts must do to resolve interpretation disputes.
The considerations relied on by the Supreme Court of Canada in resolving interpretation disputes are varied and in a given case may include the text, the purpose of the text, the context, the consequences of applying the text, appeals to legal values (such as those enshrined in the presumptions of legislative intent), and a wide range of extrinsic aids including case law, scholarly publications, legislative evolution and legislative history. However, not all these considerations are taken into account in every case. Which considerations are taken into account depends on the court's theory of interpretation. Its theory indicates the basis and limits of courts' jurisdiction in resolving interpretation disputes and suggests how this jurisdiction should be exercised. As L'Heureux-Dubé J. warns, if the Court is to give a satisfactory justification of outcomes, and one which offers adequate guidance to lower courts and other interpreters, there should be a coherence between its theory of statutory interpretation and its practice.
Probably every judge would agree that a court's first duty in resolving interpretation disputes is to give effect to the intention of the enacting legislature. Agreement would disappear, however, with the question of how judges or other interpreters can come to know the intention of the legislature. This question is complex and multifaceted. First, there is the general problem that has long plagued philosophers, namely, the problem of knowing other minds. How can a person ever really know the intentions of another? How can a person ever be sure that she knows? Second, even if knowledge of other minds were possible, legislatures don't actually have minds, so how can they form intentions? And finally, even if legislatures had minds and could form intentions, how could their intentions extend to every possible set of facts to which their legislation might apply after enactment?
The Supreme Court of Canada responds to these questions more or less directly, in several different ways. It is possible to discern in its judgments reflections of three distinct theories of statutory interpretation: (1) textualism (also called literalism), which underlies the plain meaning rule, (2) intentionalism, which underlies the modern principle of interpretation set out in the 2nd edition of Dreidger, and (3) pragmatism, which explains and justifies the actual practice of modern courts.
Textualism is built on the proposition that the only reliable indicator of legislative intention is the meaning of the legislative text. Therefore, to the extent this meaning is discernable, it should be the sole admissible evidence of legislative intent. This position is well expressed by Lamer C. J. in Ontario v. C.P. Ltd.:
[T]he first task of a court construing a statutory provision is to consider the meaning of its words in the context of the statute as a whole. If the meaning of the words when they are considered in this context is clear, there is no need for further interpretation. The basis for this general rule is that when such a plain meaning can be identified this meaning can ordinarily be said to reflect the legislature's intention.... [T]he best way for the courts to complete the task of giving effect to legislative intention is usually to assume that the legislature means what it says, when this can be clearly ascertained.(8)
The second pillar of textualism is the rule of law. For citizens to receive fair notice of the law and to arrange their affairs with reasonable security, they must be able to rely on the apparent meaning of the legislative text, which is assumed to be its literal meaning. This aspect of the plain meaning rule is emphasized in the tax cases and is also invoked on occasion in cases interpreting the Criminal Code. For example, in R. v. McIntosh, Lamer C.J. writes:
Under s. 19 of the Criminal Code, ignorance of the law is no excuse to criminal liability. Our criminal justice system presumes that everyone knows the law. Yet we can hardly sustain such a presumption if courts adopt interpretations of penal provisions which rely on the reading-in of words which do not appear on the face of the provisions. How can a citizen possibly know the law in such a circumstance?(9)
There are various strains of texualism found in North American case law, but in Canada textualism is expressed primarily through the plain meaning rule. This rule has enjoyed something of a revival in the 1990's. The key claims made by its proponents are the following:
A second response to the question of how to determine legislative intent is currently labeled intentionalism.(12) Like textualism, this theory makes the legislature's intention the primary touchstone of interpretation. But unlike textualism, it does not rely exclusively on the text of legislation for evidence of that intent. Intentionalists are willing to consider and in a proper case rely on any evidence of legislative intent so long as it meets a threshold test of relevance and reliability. This position is admirably set out by McLachlin J. dissenting in R v. McIntosh:
The point of departure for interpretation is not the "plain meaning" of the words, but the intention of the legislature. The classic statement of the "plain meaning" rule, in the Sussex Peerage Case ... makes this clear: "the only rule for the construction of Acts of Parliament is, that they should be construed according to the intent of the Parliament which passed the Act".... As Lamer C.J. put it in R. v. Z.(D.A.) ..., "the express words used by Parliament must be interpreted not only in their ordinary sense but also in the context of the scheme and purpose of the legislation". The plain meaning of the words, if such exists, is a secondary interpretative principle aimed at discerning the intention of the legislator. If the words admit of only one meaning, they may indeed "best declare the intention of the lawgiver" as suggested in the Sussex Peerage Case ..., but even here it is the intention, and not the "plain meaning", which is conclusive.(13)
McLachlin J. makes the point that for an intentionalist a formal finding of ambiguity in the text of legislation is not a prerequisite for looking at considerations other than literal meaning. The meaning of the text is important, and when it appears to be clear and precise, it should receive significant weight. But even when the text is apparently clear and precise, the court is still obliged to consider other cogent evidence of legislative intent.
This is the key difference between textualism and intentionalism. Textualists keep their eyes on the text and refuse to look at anything that might contradict the literal meaning. Intentionalists go looking for trouble. They don't always find it, of course, for in ideal circumstances, the extra-textual evidence of legislative intent supports and confirms literal meaning. But when there is a discrepancy, if the extra-textual evidence is sufficiently compelling, the intentionalist must reject literal meaning and give effect to the legislature's apparent intent.
Thus in McIntosh, McLachlin J. argues that the Court should correct an error in the legislative text to bring it in line with the evident intention of the legislature. This intention was gleaned from reading the text but also from studying the legislative history and evolution of the provision. In other cases, intentionalist courts have been led to adopt strained or implausible interpretations in order to carry out the legislature's apparent will. Often courts supplement the text by adding a qualification or exception that is not explicitly spelled out in the text. For a true literalist, adding words to a clear text is always unacceptable, for it is a form of judicial amendment. For the intentionalist, however, so long as the additional words express what there is reason to believe the legislature intended, adding words to the text is a form of interpretation not amendment.
A third response to the dilemma of determining legislative intent is known as pragmatism. Like the intentionalist, the pragmatist is interested not only in the literal meaning of the legislative text, but in the entire range of interpretive aids - purpose, consequences, extra-textual context, traditional and newly emerging legal values, and relevant extrinsic materials. Like the intentionalist, the pragmatist relies on the full range of interpretive aids in every case, not just those in which the text is found to be ambiguous. However, unlike the intentionalist, the pragmatist does not make legislative intention the sole measure of interpretation. For the intentionalist, an interpretation is correct to the extent it carries out the intention of the legislature. For the pragmatist, an interpretation is correct to the extent it solves the interpretation dispute in an appropriate way. The term "appropriate" is shorthand for a constellation of concerns. An appropriate solution must
According to pragmatists, assigning appropriate weight to these factors and balancing them against one another is the essence of the judicial function in interpretation.
While there is no consistent proponent of pragmatism on the Supreme Court of Canada, members of the Court sometimes express pragmatist views. For example, in the Régie des permis d'alcool case, L'Heureux-Dubé J. quotes and adopts a series of passages that reflect pragmatist attitudes -- from Bennion writing in the United Kingdom, from Côté and Sullivan in Canada, from Eskridge and others in the United States.(14) Her own approach is perhaps best captured in the following :
What Bennion calls the "informed inerpretation" approach is called the "modern interpretation rule" by Sullivan and "pragmatic dynamism" by Eskridge. All these approaches reject the former "plain meaning" approach. In view of the many terms now being used to refer to these approaches, I will here use the term "modern approach" to designate a synthesis of the contextual approaches that reject the "plain meaning" approach. According to this "modern approach", consideration must be given at the outset not only to the words themselves but also, inter alia, to the context, the statute's other provisions, provisions of other statutes in pari materia and the legislative history in order to correctly identify the legislature's objective. It is only after reading the provisions with all these elements in mind that a definition will be decided on.(15)
While L'Heureux-Dubé J. identifies her modern approach with the pragmatic approaches of Eskridge and others, in fact there is an important difference between her modern approach and their pragmatism. She herself draws attention to this difference when she quotes and then distances herself from the following observation by P. Michell:
At the core of [Eskridge's pragmatic] approach is a healthy scepticism about all theoretical approaches and a measure of uncertainty as to whether the answer chosen is the correct one. At the same time, however, critical pragmatism is concerned to get the job done, not to equivocate or temporize. Seen from this perspective, the essential problem of statutory intepretation is to apply a general, abstract statutory provision to a concrete factual situation. Circumstances often arise which the enacting legislation did not or could not have contemplated. Interpreters, on this account, must do what works best, by reference to the "web of beliefs" that surround a statute.(16)
L'Heureux-Dubé J. is uncomfortable with this analysis because it appears to give the judiciary carte blanche. She writes:
In my view, Eskridge's "pragmatic dynamism" provides the judiciary with a justification for manufacturing interpretations that are diametrically opposed to the clear purpose of a statute.... [The basis of Eskridge's theory] tends to diverge from the rule of law and état de droit concepts as they are accepted today in our democratic societies.(17)
Despite her discomfort, it must be acknowledged that the passage from Michell accurately reflects the spirit of pragmatism and its program. As Michell suggests, pragmatism is a response to certain intractable realities, namely, communication through natural language is never a sure thing; rules drafted by legislatures tend to be general and are often abstract; and legislatures cannot form intentions with respect to how these rules should apply to every possible set of facts. Because they do not fully address these realities, literalism and intentionalism are inadequate theories of statutory interpretation . Pragmatism is a better theory because it acknowledges these realities and attempts to reconcile them with the doctrines of parliamentary sovereignty and rule of law.
The basic premise of pragmatism is that the outcomes of interpretation disputes are appropriately determined by judges, not legislatures. The mandate of courts in resolving such disputes is not to determine the meaning of legislative texts, or even the intention of Parliament, but rather to solve the dispute in a manner that respects the important values of society. In a democracy these include deference to the elected legislature, which is heavily weighted, but also respect for the rule of law and for other well-established and emerging principles and policies belonging to the legal tradition.
This way of looking at the judicial task has a number of implications. For example:
While L'Heureux-Dubé J. may not accept the basic premise of pragmatism, she seems to endorse the corollaries that flow from that premise. Her approach to problem solving is pragmatist in many ways, but particularly in its insistence that the underlying premises of judgement must be revealed and critically assessed. As she writes in the Régie des permis d'alcool case,
given the growing recognition that there are many different perspectives -- the aboriginal perspective, for example - I believe that the era of concealed underlying premises is now over. In my view, those premises must be brought to the surface in order to promote consistency in our law and the integrity of our judicial system.(18)
L'Heureux-Dubé J. is not the only member of the Court who endorses pragmatist views. Gonthier J.'s account of statutory interpretation as set out in R. v. Nova Scotia Pharmaceutical Society(19) and elaborated in Ontario v. C.P. Ltd.(20) is also pragmatist in its recognition of the "mediating" role that judges play, not only where the language is vague or ambiguous, but in every case. In the Nova Scotia Pharmaceutical case, he points out the legislative character of this role:
... I fail to see a difference in kind between general provisions where the judiciary would assume part of the legislative role and "mechanical" provisions where the judiciary would simply apply the law. The judiciary always has a mediating role in the actualization of law, although the extent of this role may vary.(21)
This account of interpretation in the Supreme Court of Canada may suggest that at the level of theory at least, things are rather clear. Members of the court belong either to the textualist school or the intentionalist school with occasional forays into pragmatism. However, this tidy analysis is complicated by the confusing role played by Driedger's modern principle, which is cited as authority by both textualists and intentionalists. It is also complicated by the practice of the Court, which is not easily explained by its theory. Some of the more troubling aspects of this practice are examined in Part 2.
While there are many reasons to dislike the plain meaning rule, I will focus on three problems which in my view undermine the value and credibility of the rule. I will show, first, that the rule is based on a number of faulty assumptions about language and communication, second that the rule focuses too narrowly on meaning and ignores other aspects of interpretation, and third, that the rule is applied -- or not applied -- on an arbitrary basis.
Although the plain meaning rule is a legal rule, it depends on two assumptions about the role of language in communication which have little to do with law. First and foremost is the idea that the words of some texts have a "plain meaning", that "plain meaning" is not just a theoretical construct but refers to something definite in the world. Second is the assumption that it is possible to decipher the plain meaning of a text simply through the act of reading, that reading is a rule-governed, text-based procedure that yields objective results. If a text is not plain, determining the meaning requires interpretation. Unlike reading, which is objective and yields results that are the same for everyone, interpretation is subjective. Interpreters are obliged to draw inferences based on personal knowledge and to make choices based on complex and competing values. Interpretations are therefore a less reliable basis for outcomes than is plain meaning.
My purpose in this section is to test the validity of these assumptions. I will look first at the practice of the court, and then at some insights about communication based on psycholinguistic studies.
Although the rhetoric of the plain meaning rule suggests that at least some of the time communication through language is a simple and straightforward matter, in truth it is not. Determining the plain meaning of a text involves a series of steps, none of which is simple or straightforward. First, the words to be interpreted must be identified. Second, the surrounding words or "co-text"(22) must be delineated. Third, the meaning of the text must be determined having regard to the co-text and other relevant factors (if any). Fourth, the meaning must be tested for ambiguity. While these steps may be carried out together, analytically they are distinct. The examination of these steps carried out below reveals serious problems with the plain meaning rule. What purports to be plain is actually a product of uncertainty, arbitrary choice, inconsistency and confusion.
There are many "texts" in statutory interpretation. The entire body of legislation produced by a legislature constitutes a text, as do particular statutes and particular provisions, as do the judgments interpreting them. Any recorded act of communication is a text. For purposes of the plain meaning rule, the text-to-be-interpreted consists of the words whose meaning has been put at issue by an attempt to apply legislation to particular facts. Suppose, for example, that a local by-law prohibits driving a vehicle in the car pool lane of a highway unless the driver is carrying two or more passengers. Suppose that I drive a hearse in the car pool lane accompanied only by three cadavers. Everyone is likely to agree that my hearse is a vehicle and that I drove it in the car pool lane while I was carrying three of something. If disagreement arises, it will focus on whether the three cadavers are passengers within the meaning of the rule. The language on which disagreement focuses, here the word "passengers", is the text-to-be-interpreted.
The text-to-be-interpreted is normally identified by the parties early on in a dispute and it is rarely the subject of comment by a court. But given its pivotal role in applications of the plain meaning rule, this lack of attention is surprising and disturbing. At least some of the time, identifying the text involves choosing between plausible alternatives which favour different outcomes. When this happens, the outcome of the dispute ultimately depends not on the plain meaning of the text, but on the initial choice of text. To the extent they affect outcome, such choices need to be acknowledged and justified.
The potential impact of text identification on outcome is nicely illustrated in Schwartz v. Canada.(23) This involved a dispute between the Minister of National Revenue and a taxpayer over a sum received by the taxpayer as damages for the cancellation of an employment contract. A valid contract of employment had been entered into but was canceled before the taxpayer started work. The issue was whether the sum received was a retiring allowance as defined in s. 248(1) of the Income Tax Act:
"Retiring allowance" means an amount ... received ... in respect of a loss of an office or employment.
"Employment" means the position of an individual in the service of some other person.
The Court found that the sum received by the taxpayer in this case was not a retiring allowance because the taxpayer was not an individual "in the service of" another and therefore did not have employment within the meaning of the section. Speaking for the Court on this point La Forest J. wrote:
The key element in the words chosen by Parliament to deal with this situation is the definition of "employment" which is the "position of an individual in the service of some other person". The statutory requirement that one must be "in the service" of another person to be characterized as an "employee" excludes, in my opinion, any notion of prospective employment when the phrase is given its ordinary meaning.... During oral argument, counsel admitted that an ordinary person would find that Mr. Schwartz was not an employee of Dynacare when the contract was canceled.(24)
The Court here identifies "employment" as the text to be interpreted so that, for the Minister to win, he must establish that "employment" means "prospective or current employment". The Court was not persuaded.
The outcome might have been different, however, if the Court had identified "loss" or "loss of employment" as the text. On this approach, the Minister could persuasively argue that "loss" means "being deprived of something a person has or is entitled to have".(25) Although the taxpayer's employment had not begun, it would have begun were it not for the cancellation. What the taxpayer lost, what he otherwise would have had, was precisely the position of being in the service of Dynacare and the benefits that would flow from that position. In fact, this is what is always lost when a person suffers a "loss of employment" - the entitlement to perform work for reward in the future. While an ordinary person would say that Mr. Schwartz was not an employee of Dynacare when the contract was canceled, that same person would also say that Mr. Schwartz lost his employment with Dynacare when the contract was canceled. There is no contradiction here because "loss" includes being deprived of things one was going to receive as well as things received.
My point is not to suggest that Schwartz was wrongly decided (though in my view it was) but to show that the outcome was not dictated by ordinary meaning; it turned in part at least on the choice of text. To focus on "employment" rather than "loss" or "loss of employment" was a choice -- in the circumstances a significant choice. It must be acknowledged and justified if the Court's interpretation is to be persuasive.(26)
First impression meaning is the meaning that spontaneously occurs to a reader upon initial reading of a text. The notion of a first impression meaning that is either plain or ambiguous lies at the heart of the plain meaning rule. This notion involves distinguishing between reading and interpretation, between internal and external context, between things that can be relied on to determine initial meaning and things that can be relied on only to resolve ambiguity. However the distinction is drawn, the basic idea is the same. First you determine the meaning of the text through reading alone, and if the meaning is plain your task is done.
The "co-text" is the portion of surrounding text that is taken into account during the initial determination of meaning. As Jacob Mey points out, linguists disagree on how much of a surrounding text should be included.(27) A common approach, however, is to include in the co-text as much of the surrounding text as the reader manages to hold in her short term or working memory.(28) For purposes of statutory interpretation, this would normally include the section or subsection in which a text appears; it might include a series of related sections. It would include at least as much of the surrounding legislation as is required to make sense of the text to be interpreted.
Notice that the co-text differs from other aspects of context in being limited to the words on the page. An interpreter who is a strict textualist would expect to derive first impression meaning from reading the text in light of the co-text alone, ignoring extra-textual features like purpose or consequences. An interpreter who is not a strict textualist would expect first impression meaning to depend not only on the text and co-text but also on whatever knowledge of context she brings to the text.
Judges often overlook the problem of identifying an appropriate co-text. Particularly among proponents of the plain meaning rule, there is a tendency either to deny the relevance of the co-text altogether, or to enlarge it so that it includes the entire statute. In R. v. McCraw,(29) for example, the issue was whether a threat to rape a young woman was contrary to s. 264.1 of the Criminal Code, which made it an offence to utter a "threat to cause death or serious bodily harm to any person". No one doubted that the young woman was a "person" or that she'd received a "threat to cause" something, although not a threat to cause "death". The doubtful question was whether she'd received a threat to cause "serious bodily harm". These words thus constituted the text to be interpreted. Their initial meaning was purportedly determined by looking at the definition of "bodily harm" in the Code and the definition of "serious" in the dictionary. Speaking for the Court, Cory J. wrote:
The appellant urged that serious bodily harm is ejusdem generis with death. I cannot accept that contention. The principle of ejusdem generis has no application to this case. It is well settled that words contained in a statute are to be given their ordinary meaning. Other principles of statutory interpretation only come into play where the words sought to be defined are ambiguous. The words "serious bodily harm" are not in any way ambiguous.(30)
Cory J. implies that to determine the plain meaning of the text "serious bodily harm", these words must be read in isolation, ignoring the rest of the sentence in which they appear or at least not letting the rest of the sentence affect the reader's understanding. But as every reader knows, this is not how reading works. Cory J.'s analysis may be legitimate, but it is not reading.
With Ontario v. C. P. Ltd., we go to the other extreme. One issue in that case was how to interpret the words "for any use" in s. 13(1) of Ontario's Environmental Protection Act. Lamer C.J. wrote:
... the first task of a court construing a statutory provision is to consider the meaning of its words in the context of the statute as a whole. If the meaning of the words when they are considered in this context is clear, there is no need for further interpretation.(31)
In order to apply this approach in the present case, it is first necessary to determine whether the terms of s. 13(1)(a) have a "plain meaning" when viewed in the context of the statute as a whole.... Although the word "use" is somewhat ambiguous when considered on its own, the expression "for any use that can be made of [the natural environment]" has, in my view, an identifiable literal or "plain" meaning when viewed in the context of the E.P.A. as a whole, particularly the other subsections of s. 13(1).(32)
Presumably Lamer C.J. does not mean to claim that he read the whole of the Environmental Protection Act before he formed an impression of the meaning of "use" in s. 13(1). Once again, that is not how reading works.(33) To consider a provision "in the context of the statute as a whole", unless the statute is short, normally you do not read it from start to finish the way you might read a newspaper article or a report. The usual thing is to read the provision first and on this basis to form an impression of what it means. You then skim through the statute, looking for related sections, for other instances of relevant words, for patterns and variations of patterns. You read other provisions of the statute selectively. You also try to figure out the structure of the Act by looking at the sequence of marginal notes, noting how sections are grouped together under headings and titles, working out the relationships among the parts and seeing how they function together to form a workable scheme. That is what is usually meant by considering or viewing a legislative text in the context of the statute as a whole. But if that is what Lamer C.J. means, he is no longer talking about first impression meaning or even literal meaning. With this sort of analysis, the distinction between textualism and intentionalism has partly broken down.
Once again, my point is not to suggest that these cases were wrongly decided, but merely to draw attention to the significant choices involved in determining plain meaning. In making this determination, the scope of the co-text obviously matters. In the C.P. case, Lamer C.J. says that the word "use" is somewhat ambiguous when considered on its own, but is plain when considered in the co-text of s. 13(1).(34) In the McCraw case, Cory J. finds the expression "serious bodily harm" to be plain when considered on its own, but apparently less plain, or perhaps different, when juxtaposed to "death".(35) The scope of the co-text matters, but as these examples show, it is a variable that is easily manipulated by the courts.
One of the most frustrating aspects of the plain meaning rule is trying to understand what sort of meaning interpreters have in mind when they label a meaning plain. There is a rich and shifting set of terms associated with plain meaning -- ordinary meaning, literal meaning, common sense meaning, ordinary and grammatical sense, natural sense, and the like. These terms have no fixed or precise reference. Sometimes they are used as synonyms for "plain meaning" but it is also clear that different judges mean different things by them.
The problem is compounded by the complexity of the subject. There are many different senses of "meaning", which linguists take care to distinguish but which lawyers and judges tend to use indiscriminately. To appreciate the extent of the confusion, it may be helpful to note the different senses of meaning that are routinely referred to in judgments under a variety of different names.
First, there is the "dictionary" meaning of words. Dictionary meaning is a-contextual word meaning. It is an attempt (always incomplete) to describe the customary ways in which individual words may be used. Dictionary descriptions are abstractions from the actual or imagined contexts in which the words actually occur. If you want to know the dictionary meaning of a word, you can look it up, or alternatively you can make it up. The courts assume that the meaning of words known by every competent user of language is equivalent to their dictionary meaning.(36)
Second, there is "literal" or "facial" meaning. Literal meaning is a-contextual sentence meaning. Like dictionary meaning, it is an abstraction; but unlike dictionary meaning it an abstraction from a particular context rather than a wide range of possible contexts. A-contextual literal meaning is normally contrasted with context-dependent "utterance" meaning. In Introduction to Natural Language Semantics, Henriette de Swart explains the distinction as follows:
[W]e can say that there are three essential ingredients to the use of language as a means of communication, namely:(i) the linguistic expression(s) used(ii) what the expression refers to (objects, properties, relations, events, ...)(iii) context.(37)
Semantic research focusses on the relation between (i) and (ii), in particular on the meaning which arises out of the combination of more elementary expressions into groups of words and sentences.(38)
The relation between (i) and (ii) yields literal meaning, the meaning studied by semanticists. As de Swart explains, they devise formal rules which describe how to organize elementary expressions into sentences and how to extract meaning from sentences and sentence parts. If you want to know the literal meaning of a text, you can apply these rules. However, like dictionaries, the rules devised by semanticists are hypothetical models of meaning formation; they do not purport to describe the actual process by which meanings are formed.
The third type of meaning is "utterance" meaning, more commonly called "speaker's" meaning or "intended" meaning. It is the type of meaning that Dreidger's modern principle is designed to yield. Utterance meaning depends on all the elements of communication in de Swart's list -- the linguistic expressions used, what they refer to, and the context. Context consists of all the knowledge that is stored in the minds of writers and readers. Assuming this knowledge is shared, it permits writers to predict readers' interpretations of the words they write and it permits readers to reconstruct writers' intentions in writing those words. Context includes not only knowledge of subject matter but also knowledge of the situation in which the text was produced and the conventions of that type of writing. If you want to know the utterance meaning of a statutory provision, you have to look at the relevant conventions of statute law, the likely purpose of the statute, the scheme devised to realize that purpose, and so on.
A fourth type of meaning is audience-based "ordinary" meaning. This is the meaning that would be understood by the members of a particular audience - whether the public at large or a particular portion of the public. When the audience is taken to be the public at large, audience-based meaning is still called "ordinary" meaning. When the audience is a specialized sub-group, audience-based meaning is usually called "technical" meaning. When the audience is a specialized sub-group consisting of lawyers and judges, audience-based meaning is actually "legal" meaning, though it is not necessarily recognized as such by lawyers and judges for whom such meanings have become as natural and ordinary as air.
"Ordinary" meaning is usually taken to be first impression meaning, the meaning that would come to the mind of the intended reader upon reading a text in ordinary circumstances.(39) Because it is first impression meaning, it is formed with reference to a limited co-text; but unlike literal meaning, ordinary meaning draws on a rich context consisting of all the knowledge that is stored in the reader's mind and accessed upon reading the text. If you want to know the ordinary meaning of a provision, strictly speaking you should conduct a survey of the relevant audience.(40) But if there's no time for a survey, you may simply speculate on how that audience would understand the text to be interpreted.(41)
Notice that, as defined here, ordinary meaning is not the same as dictionary meaning or literal meaning; ordinary meaning is contextual whereas dictionary and literal meanings are not. Notice, too, that ordinary meaning differs from utterance meaning in an important respect. In the case of utterance meaning, the touchstone of successful communication is the speaker's intended meaning which must be inferred by the reader with reference to the speaker's imagined context; in the case of ordinary meaning, it is the meaning understood by the reader with reference to her own context.
Finally, there is the concept of "applied" meaning-- the meaning of the text in relation to particular facts. Unlike linguistics, statutory interpretation is not an academic exercise; the point is not to study how communication occurs but to answer a question or resolve a dispute. If you want to know the applied meaning of a provision, you must determine the meaning of the legislation in relation to particular facts. You must determine whether the rule and applies, and if so, what the legal implications are.(42)
When judges refer to the ordinary or literal meaning of a text or when they invoke the plain meaning rule, it is often unclear which of these different types of meaning they have in mind. And when judges are clear about what they mean, it turns out that they often mean different things by these terms. When Lamer C.J. uses the term "plain meaning" in Ontario v C.P. Ltd. he means plain literal meaning.(43) When Cory J. uses it in the Alberta (Treasury Branches) v. Canada, he seems to mean plain applied meaning.(44) When L'Heureux-Dubé J. uses it in Manulife Bank of Canada v. Conlin, she means plain ordinary (or audience based) meaning.(45) LaForest J. generally avoids the term "plain meaning" but he often refers to "ordinary meaning" by which he usually means dictionary meaning,(46) but sometimes audience based meaning as defined above.(47)
In Friessen v. Canada, Major J. identifies "plain meaning" with the meaning established by the common law.(48)
There is more at stake here than confusing or inconsistent terminology. These terminological problems reflect genuine uncertainty and inconsistency in the Court's understanding of what is meant by plain meaning and how it is established. If the Court wants to take a literalist position, then strictly speaking the only kind of meaning that should interest it, initially at least , is literal meaning; and the only context it should look at is the co-text. Purpose and consequences, the Act as a whole, legislative policies, extrinsic aids -- none of this should be relevant at stage one, when a court through reading alone determines the plain meaning of the text.
Sometimes members of the Court are uncompromisingly literal. In recent judgments Lamer C.J., for example, has taken a strong literalist stance. In R. v. McIntosh, the Court was concerned with the application of ss. 34 - 37 of the Criminal Code dealing with self-defence. The defendant in McIntosh had killed to defend himself from an assault which he himself had provoked. The issue was whether he was obliged to meet the onerous conditions for self-defence set out in s. 35 or could rely on the less onerous conditions of s. 34(2).
[Self-defence against unprovoked assault]
34.(1) Every one who is unlawfully assaulted without having provoked the assault is justified in repelling force by force if ...
[Extent of justification]
(2) Every one who is unlawfully assaulted and who causes death or grievous bodily harm in repelling the assault is justified if ...
[Self-defence in case of aggression]
35. Every one who ... has without justification provoked an assault on himself by another, may justify the use of force subsequent to the assault if ...
The text to be interpreted was the opening words of s.34(2). The co-text was the remainder of the subsection. Speaking for the majority, Lamer C.J. wrote:
While s. 34(1) includes the statement "without having provoked the assault", s. 34(2) does not. Section 34 (2) is clear, and I fail to see how anyone could conclude that it is, on its face, ambiguous in any way. Therefore, taking s. 34(2) in isolation, it is clearly available to an initial aggressor.
The Crown has asked this Court to read into s. 34(2) the words "without having provoked the assault". The Crown submits that by taking into consideration the common law of self-defence, legislative history, related Criminal Code provisions, margin notes, and public policy, it becomes clear that Parliament could not have intended s. 34 (2) to be available to initial aggressors.(49)
In this passage Lamer C.J. refuses to look at any extra-textual features of context; he refuses to look at marginal notes or the other provisions of the Act. Elsewhere in the judgment he acknowledges these features and comments on them, but he does not let them influence his reading of the text, which is based on literal meaning alone.
Other members of the Court have also taken a strong literalist stance, particularly in judgments interpreting the Income Tax Act. The following passage is from the judgment of Iacobucci J., speaking for the Court in Canada v. Antosko:
While it is true that the courts must view discrete sections of the Income Tax Act in light of the other provisions of the Act and of the purpose of the legislation, and that they must analyze a given transaction in the context of economic and commercial reality, such techniques cannot alter the result where the words of the statute are clear and plain and where the legal and practical effect of the transaction is undisputed....(50)
This is a clear and coherent exposition of the literalist position: other provisions of the Act, legislative purpose, and real world consequences have no role to play in the initial determination of meaning.
Although members of the Court frequently invoke the plain meaning rule, they do not always adhere to the literalist position that is supposed to go with it. In the Court's recent case law, sometimes the plain meaning of texts is determined having regard to the legislature's purpose, sometimes not. Sometimes the presumptions of legislative intent are relied on, sometimes not. Sometimes the absurd consequences of an interpretation are taken into account, sometimes not.
The Court's understanding of the role of purpose in determining meaning is a particularly striking illustration of this inconsistency in approach. In Canada v. Friesen, Major J. emphasizes the irrelevance of purpose:
... the clear language of the Income Tax Act takes precedence over a court's view of the object and purpose of a provision....
Therefore, the object and purpose of a provision need only be resorted to when the statutory language admits of some doubt or ambiguity.(51)
In other cases, however, proponents of the plain meaning rule have been less clear about the role of purpose. In R. v. Adams, for example, Sopinka J. (speaking for a bench that included Lamer C.J., Iacobucci J., and Major J.) writes:
In approaching the interpretation of any statutory provision, it is prudent to keep in mind the simple but fundamental instruction offered by the court in Reigate Rural District Council v. Sutton District Water Co ... and affirmed by this Court in Hirsch v. Protestant Board of School Commissioners ...:
...it is always necessary in construing a statute, and in dealing with the words you find in it, to consider the object with which the statute was passed, because it enables one to understand the meaning of the words introduced into the enactment.
This well-settled rule of statutory interpretation has continued to be followed by this Court to the present time.(52)
On this analysis, purpose must be taken into account at the outset, to determine what the provision means and whether the meaning is plain. A similar approach is adopted by McLachlin J. in Opetchesaht Indian Band v. Canada. She begins her review of the principle governing statutory interpretation with the following observation:
This court has recently affirmed that the process of statutory interpretation requires that the intention of Parliament be ascertained first by considering the plain meaning of the words used in the statute, and has determined that where "the words used in a statute are clear and unambiguous, no further step is needed to identify the intention of Parliament"....(53)
She ends, however, by insisting that plain meaning cannot be equated with dictionary meaning or other forms of a-contextual meaning. At the very least, purpose must be taken into account to determine whether the meaning is plain:
... as the principles of construction explored above suggest, dictionary or "plain" meanings suffice only where they are clear and consistent with a purposive reading of the statute as a whole....When read in the context of the purpose of the Act, what seems at first blush to be a "plain meaning" may be revealed as not so plain after all.(54)
The conviction that purpose must be looked at from the outset is found in tax cases too. In Corporation Notre-Dame de Bon Secours v. Communauté de Québec, for example, the Court adopts a teleological or purposive approach and asserts that "the first consideration [in interpreting tax legislation] should ... be to determine the purpose of the legislation, whether as a whole or as expressed in a particular provision." (55)
The confusion over the role of purpose in determining meaning is nicely captured in the following passage from the judgment of Cory J. in Alberta (Treasury Branches) v. Canada.
... [W]hen there is neither any doubt as to the meaning of the legislation nor any ambiguity in its application to the facts then the statutory provisions must be applied regardless of its object or purpose.... [In this case] neither the meaning of the legislation nor its application to the facts is clear. It would therefore seem to be appropriate to consider the object and purpose of the legislation. Even if the ambiguity were not apparent, it is significant that in order to determine the clear and plain meaning of the statute it is always appropriate to consider the "scheme of the Act, the object of the Act, and the intention of Parliament."(56)
The first sentence of this passage says that purpose is irrelevant when the meaning is plain. The final sentence says that we should always look at the purpose to determine the plain meaning. The Court apparently finds this analysis helpful, for it has been quoted with approval in subsequent cases.(57)
The way the Court deals with the problem of absurd consequences is equally unsatisfactory. In some cases we are told that absurdity and injustice are irrelevant in determining the plain meaning of a text. For example, in R. v. McIntosh, Lamer C.J. writes:
... where, by the use of clear and unequivocal language..., anything is enacted by the legislature, it must be enforced however harsh or absurd or contrary to common sense the result may be. The fact that a provision gives rise to absurd results is not, in my opinion, sufficient to declare it ambiguous and then embark upon a broad-ranging interpretive analysis.(58)
In other cases, however, the fact that clear and unequivocal language gives rise to absurd results is sufficient reason to look long and hard for an alternative interpretation.
In R. v. Boma Manufacturing,(59) for example, the Court was required to interpret s. 165(3) of the Bills of Exchange Act, which provided that "where a cheque is delivered to a bank for deposit to the credit of a person and the bank credits him with the amount of the cheque, the bank acquires all the rights and powers of a holder in due course of the cheque." Although cheques had undoubtedly been delivered to the respondent bank for deposit to the credit of a person whose account was in fact credited, the Court nevertheless denied the bank the rights and powers of a holder in due course. Iacobucci J. (speaking for Lamer C.J. and others) wrote:
The respondent submits that, within the plain meaning of s.165(3), it has acquired the rights of a holder in due course, since the cheques in question were indeed "delivered to a bank for deposit to the credit of a person" and since the CIBC credited the person "with the amount of the cheque". At first blush, this interpretation seems to be attractive. However, the consequence of this approach would be far-reaching and overly broad.
If the respondent's interpretation were adopted, a bank would never need to require an endorsement.... A bank would always be immune from the consequences of having accepted unendorsed cheques into third party accounts. This result cannot be supported.(60)
The Court here declines to find that the meaning of the provision is plain because it leads to unacceptable consequences. It does exactly what is said in R. v. McIntosh to be impermissible: it treats the word "person" is if it were ambiguous, embarks on an interpretive analysis, and ends by rewriting the text. It concludes, largely on basis of the provision's purpose, that "person" in s. 165(3) actually means "person who is entitled to the cheque".(61) In other words, it adds words to the provision which were not actually there.
In all these cases the Court consistently asks the same question: does the text have a plain meaning? However, in answering this question it does not consistently rely on the same sense of "meaning". In some cases it focuses on literal or a-contextual meaning; in other cases selected aspects of the context are taken into account -- sometimes purpose only, sometimes consequences only, sometimes other statutes or the common law. No explanations or justifications are offered for these inconsistencies. In fact, it appears that most of the time they are not even noticed. It is difficult to resist the suspicion that, without being aware of it, the Court takes into account as much context as it needs to support its preferred interpretation, declares that interpretation to be the meaning plain of the text, and dismisses all other contextual features as irrelevant. If this is so, plain meaning is not dictating the outcome; rather the outcome (preferred on other grounds) is dictating plain meaning. This approach to statutory interpretation is objectionable not because judges have a preferred outcome,(62) but because the real determinants of the outcome remain hidden and the proffered explanation is unpersuasive.
The distinction between "plain" and "ambiguous" lies at the heart of the plain meaning rule. In fact, everything turns on it. If a text is plain, there is no need for interpretation; the plain meaning prevails over other evidence of legislative intent to the extent of any discrepancy, so that the other evidence becomes irrelevant. But if the text is ambiguous, interpretation is required and that other evidence must be relied on to resolve the ambiguity. Since the distinction between plain and ambiguous is central, one would expect a good deal of attention to be given to these concepts and to developing appropriate tests and procedures for telling them apart. In fact, little attention is paid to this aspect of the rule.
For purposes of the plain meaning rule, a text is said to be ambiguous if it is reasonably capable of bearing more than one plausible meaning. This definition encompasses the two standard forms of lexical ambiguity, namely, semantic ambiguity (a word has two or more senses) and syntactic ambiguity (the structure of a sentence can be read in two or more ways). But what about other forms of linguistic imperfection -- like vagueness or incoherence? Do these count as ambiguity for purposes of the rule? The answer is that sometimes these are sources of ambiguity, but when it suits the Court's purpose they are not.
In the McIntosh case, for example, Lamer C.J. begins his analysis of s. 34(2) of the Criminal Code by observing that " the conflict between ss. 34 and 35 is obvious on the face of the provisions."(64) In the next paragraph he notes that the provisions are "highly technical, excessively detailed ... and are internally inconsistent in certain respects. Moreover, their relationship to s. 37 ... is unclear."(65) He ends his critique by agreeing with the trial judge that "these sections of the Criminal Code are unbelievably confusing."(66) But having said all that, Lamer C.J. goes on to find that the meaning of s. 34(2) is plain. He fails to see "how anyone could conclude that it is, on its face, ambiguous in any way."(67) In this case at least, the concept of ambiguity is a narrow one -- one that allows a text to be both unbelievably confusing and obviously plain.
Another disturbing aspect of the distinction between plain and ambiguous is the Court's reliance on the concept of plausibility. Every text permits multiple readings that are more or less plausible. Psycholinguists point out that the shorter a text and the more limited its co-text and context, the greater the scope for multiple readings.(68) While competent language users are likely to agree about some readings, they are almost sure to disagree about others. There is no firm line between plausible and implausible readings, no principled way to determine the point at which a proposed reading may or must be dismissed as implausible. Because judgements about plausibility are rooted in the sensibilities of individual speakers, they are highly subjective and naturally differ from one reader to the next. So when judgements differ, whose sensibility governs? And if the matter is really a linguistic one, why do we let judges decide? Wouldn't it make more sense to appeal to the expertise of linguists?(69)
Another troubling aspect of the plain meaning rule is its proponents' disregard of actual disagreement about the meaning. A text has a plain meaning if it can plausibly be read in only one way. Given this, a claim made by another judge to read the text in a different way should be good evidence that the text does not have a plain meaning. Yet such claims are routinely ignored by proponents of the plain meaning rule.(70)
Consider, for example, the judgments of Lamer C.J. and Gonthier J. in Ontario v. C.P. Ltd. Both define their task as determining the meaning of the words "any use that can be made of it" in the following provision of Ontario's Environmental Protection Act:
13.(1) ... no person shall deposit ... a contaminant ... into the natural environment ...
(a) that causes or is likely to cause impairment of the quality of the natural environment for any use that can be made of it ....
Both consider the meaning of these words in a co-text consisting of the seven other clauses, (b) through (h), comprising the subsection. After analyzing the language of these clauses, Lamer C.J. concludes:
When these factors are taken into account, it can, I believe, be concluded that the literal meaning of the expression "for any use that can be made of [the natural environment]" is "any use that can conceivably be made of the natural environment by any person or other living creature".(71)
However, Gonthier J., considering the same text and analyzing it in the same co-text, reaches a different conclusion. He writes:
The choice of terms in s. 13(1) leads me to conclude that polluting conduct is only prohibited if it has the potential to impair a use of the natural environment in a manner which is more than trivial.(72)
In other words, the meaning of the expression "for any use that can be made of [the natural environment]" is "for any non-trivial use ...".
These judges come up with different meanings because they notice different things. Lamer C.J. notices that elsewhere in s. 13(1), the word "use" was qualified and he inferred that since it was not qualified in s. 13(1)(a), its meaning there must be "any use whatsoever, without qualification". Gonthier J. notices that the other uses listed in clauses (b) through (h) all have a potential to cause significant injury or damage to the environment and inferred that "use" in clause 13(1)(a) must be similarly limited.
Upon completing his analysis of s. 13(1) , Lamer C.J. declares that the literal meaning of the text is plain rather than ambiguous so that, absent constitutional considerations, this meaning must govern. In other words, in his view the text is not capable of bearing two plausible meanings. This conclusion entails that the meaning identified and preferred by Gonthier J. is not plausible. But on what basis does Lamer C.J. not only prefer his reading to Gonthier J.'s, but dismiss Gonthier J.'s as implausible? Is Gonthier J. not a competent user of language? Are his linguistic intuitions not equal to the Chief Justice's? Obviously Lamer C.J. is not claiming to have superior linguistic skills, but that is the logical implication of his reliance on the plain meaning rule in these circumstances. In a fundamental way, the rule is insulting to those who disagree, for it dismisses their arguments rather than answering them. The competing interpretation is dismissed as being linguistically implausible and the arguments in favour of that interpretation are dismissed as legally irrelevant.
At first blush, the claim that some texts have a plain meaning seems clear and easy to test. However, a look at recent case law shows that in practice determining the plain meaning of a text depends on an arbitrary identification of the text and co-text, shifting and uncertain conceptions of meaning, and highly subjective judgements about plausibility. In my view, these problems undermine the credibility of the plain meaning rule and in particular the claim that if the meaning is clear, the outcome is determined not by the judge but the text. In reality, judicial discretion is not taken away by plain language; it would be more accurate to say that it is driven underground. The outcome is the result of considerations that are not acknowledged, possibly not noticed, and certainly not justified to any acceptable degree.
Textualists argue that if legislation has a plain meaning, then courts are constitutionally obliged to adopt it, for any departure from the plain meaning would wrongfully amend the legislation and usurp legislative power. This claim obviously depends on the assumption that plain meaning is the same for everyone so that if a text is plain, the meaning understood by the court can safely be equated with the meaning intended by the legislature as well as the meaning understood by the audience to which the legislation is addressed.
This view of language as a more or less transparent carrier of meaning is widespread in our culture. If a writer's thoughts are correctly embodied, by using the right words arranged in the right order, the meaning will be communicated to the reader automatically simply through reading. As Georgia Green points out in Pragmatics and Natural Language Understanding, this sanguine view of communication has been dubbed "the conduit metaphor":
According to the conduit metaphor, linguistic expressions (words, sentences, paragraphs, books, etc.) are compared to vessels or conduits into which thoughts, ideas, or meanings are poured, and from which they can be extracted, exactly as they were sent, accomplishing a transfer of possession.... When we accept the conduit metaphor, we commit ourselves to a view that communication is achieved as easily as serving a glass of milk or sending a package, than any failure to communicate must be due to carelessness or inattention in choosing or construing linguistic expressions, and that properly chosen linguistic expressions do all the work.(73)
In Green's view, this conception of communication is mistaken and misleading. In place of the conduit metaphor she suggests, following Reddy,(74) that linguistic expressions are actually like blueprints - "blueprints from which much may be inferred, but with no assurance of success." The blueprint metaphor is appropriate because it emphasizes the ambiguous and indeterminate nature of the language that comprises texts and the extensive work that readers must do to infer intended meaning. Green concludes:
.... there is more to understanding utterances than parsing them and deriving representations of their propositional meanings.... It is necessary also to make inferences about what the utterer believes about what the addressee believes, and about what effect the utterer intends the utterance to have.(75)
It is evident from this account that the conduit metaphor conceives of meaning as literal meaning, the meaning studied by semanticists. The blueprint metaphor, on the other hand, conceives of meaning as utterance meaning, the sort studied by pragmatists. It is also evident that when textualists assert or assume that plain meaning is the same for all and therefore equivalent to the meaning intended by Parliament, they are relying on the conduit metaphor.
This recognition is important for several reasons. First, it underlines that plain meaning is, and must be, conceived of as literal meaning -- the kind of meaning delivered "through" the text, independently of readers' knowledge and inferences. Plain meaning can be considered objective and universal only in so far as it is derived from a stable lexicon and a fixed syntax that is the same for all competent language users. We have seen that judges sometimes fudge this point, bringing purpose or other aspects of context into their determination of plain meaning. But in principle plain meaning is literal meaning, as Lamer C.J. assumes in his judgments.
The recognition that the plain meaning rule is based on the conduit metaphor is important because it shows that the rule is tied to an understanding of communication that is no longer credible. In recent years, primarily as a result of unsuccessful efforts to get computers to read and write, linguists have discovered the pervasive ambiguity of word meaning and syntax. In Lexical Ambiguity Resolution: Perspectives from Psycholinguistics, Neuropsychology & Artificial Intelligence, Prather and Swinney write:
Ambiguity is ubiquitous in language; it exists at every level of processing (from acoustic/ phonetic to semantic to structural, etc.).(76)
In the same book Simpson and Burgess write:
[A]mbiguity arising from the fact that some lexical items have two or more distinct dictionary entries, is simply an extreme and obvious example of a general vagueness or indeterminacy that is pervasive in language... Research ... suggest [sic] that all words carry with them more information than a person requires for comprehension of any particular message. The processes by which some of this information is selected for use in context are no less important for single-meaning words than they are for ambiguous ones....
This indeterminacy of meaning by no means stops at the lexical level. Ambiguity at the syntactic level has generated considerable research as well....(77)
Notice that the claim made by these linguists is not that communication is impossible, but rather that language, taken by itself, is not enough; something more is required to resolve the ambiguity. In A Theory of Textuality: The Logic and Epistemology, Jorge Gracia writes:
... texts are always given in a certain language that obeys rules and whose signs denote and connote more or less established meanings. In addition, the audience cannot help but bring to the text its own cultural, psychological, and conceptual context. Indeed, the understanding of the meaning of a text can be carried out only by bringing something to the text that is not already there....(78)
As Frank Smith explains,
This complicated ambiguity of language is the reason that it is difficult to program computers to translate language or make abstracts, even when they are equipped with a "dictionary" and a "grammar". Computers lack the knowledge of the world that is required to make sense of language.(79)
Communication, it turns out, requires nothing less than an encyclopaedic knowledge of the world. This includes not only specific knowledge of the subject matter of the communication, the "scenes" and "scenarios" invoked, (80) the "genre" and its associated conventions(81) but also the vast body of assumptions, beliefs, opinions, and values that the reader takes to be true or correct. François Rastier points out that to appreciate the meaning of a text, even the simplest of texts, "one must often have recourse to encyclopaedic knowledge stemming not only from the social sciences but also from the natural sciences."(82) It is this knowledge that supplies context and permits lexical ambiguity to be resolved or at least brought to a manageable level.
The pervasive indeterminacy of language is rarely noticed by readers. Most of the inferences required to resolve doubts about the meaning of a text are made quickly and easily, at a subconscious level. The problem is that among competent users of language these questions and doubts are not necessarily resolved in the same way. To draw the same inferences, writers and readers must share roughly the same knowledge, and similar if not identical values. In Semantic Theory Ruth Kempson writes:
... in analysing communication, a large number of indeterminacies arise since the content of the communication may vary according as the assumptions of the particular speaker and particular hearer vary.(83)
In Discourse Analysis, Gillian Brown and George Yule write:
However objective the notion of "text" may appear as we have defined it (the verbal record of a communicative act), the perception and interpretation of each text is essentially subjective. Different individuals pay attention to different aspects of texts. The content of the text appeals to them or fits into their experience differently. In discussing texts we idealise away from this variability of the experiencing of the text and assume what Schutz has called 'the reciprocity of perspective', whereby we take it for granted that readers of a text ... share the same experience (Schutz, 1953).(84)
The assumption of reciprocity is necessary for efficient social interactions, but it is an assumption that is not necessarily borne out. The larger and more diverse the community, the greater the cultural distance among different readers, the less likely they are to share context and the greater the likelihood is -- not that readers will experience a text as being ambiguous - but rather that readers will experience plain, but different, meanings. As Frank Smith observes, readers usually see what they are looking for in a text and remain unaware of the other possibilities.(85)
It is not only pragmatists who insist on the impossibility of communicating without context. Semanticists make the same point. In Introduction to Natural Language Semantics, Henriette de Swart writes:
Any act of communication takes place in a specific context, so communication always relies on utterance meaning. On the other hand, our linguistic capacity is clearly independent of any specific context in which we utter a sentence.(86)
Semanticists study linguistic competence, as opposed to actual linguistic performances; they study an abstraction from reality rather than the thing itself. As Francois Rastier explains, literal meaning (or "signification" as he calls it) is not something to which contextual meaning is added, like warm clothing on a cold day:
Meaning is not added to a signification that is already there. On the contrary, signification results from an abstraction carried out by the linguists starting from meaning.... Signification ... is an artefact created by linguists and it remains inevitably equivocal.(87)
Jorge Gracia writes:
What needs to be stressed is that the meanings of all texts depend on context to some extent. This entails that there is no such thing as the "literal" meaning of a text if by 'literal meaning' is understood meaning apart from context.(88)
If texts must be read in a context in order to be meaningful, if literal meaning is merely a theoretical construct, then there is no such thing as plain meaning. If texts are always read in a context supplied by the reader, then reading is not different in kind from interpretation. Like interpretation, reading is a creative, subjective activity. Here is how Frank Smith defines it in Understanding Reading: A Psycholinguistic Analysis of Reading and Learning to Read:
Reading is never an abstract, purposeless activity, although it is frequently studied in that way.... Readers always read something, they read for a purpose, and reading and its recollection always involve feelings as well as knowledge and experience. In other words, reading can never be separated from the purposes of readers and from its consequences upon them.(89)
Reading is ... a creative and constructive activity having four distinctive and fundamental characteristics - it is purposeful, selective, anticipatory, and based on comprehension, all matters where the reader must clearly exercise control. (90)
Readers can derive meaning directly from text because they bring expectations about meaning to a text. Prediction, based on prior knowledge, eliminates unlikely alternatives in advance.(91)
The crucial prerequisite of reading is prior knowledge. The knowledge a reader brings to a text is relied on to draw inferences, make guesses and predictions, and eliminate implausible possibilities. In short, it is relied on to determine the meaning of the text.
If there is no such thing as plain meaning, then what are judges referring to when they purport to ignore context and rely only on the literal meaning of a text, with or without co-text? The answer is that in fact they are not excluding context; they are not relying on dictionary definitions and grammar rules. They actually are relying on context because it is impossible to read without it. As Georgia Green explains, there are no "null" contexts in which utterances can be interpreted:
As Crain and Steedman put it, "the so-called null context is simply an unknown context."(92) When we are asked to act as informants, and make judgements about expressions or their meanings "out of context" or "in a null context", we cannot help but imagine SOME context.... We differ, as individuals, and on occasions, in how much context we import into the judgement task, and in what we are willing to imagine when we try to construe the expression as a sensible thing to utter on an occasion of the sort we assume.(93)
All of us always bring something to the text. It is not possible for judges who interpret a provision of the Criminal Code or the Income Tax Act to wipe out the beliefs, values and expectations that they bring to their reading. They cannot erase their knowledge of law or of the subject of legislation. They cannot cast aside legal culture, with its respect for common law and evolving constitutional values. They cannot unexperience their experiences, unread the books they have read, unwatch the television they have seen. All this necessarily informs judicial notions of what is true, normal, reasonable, plausible, desirable and fair -- notions that are essential in constructing meaning. Like any other readers, if they want to make sense of a text, judges must rely on the context that they themselves bring to the text.
Because statutes are in the form of texts, a quasi-physical form, they help create an illusion that the law is an artifact, that is, a fully determined, pre-existing thing. It is easy to confound the text -- the written record of an Act of Parliament -- with the law. But a text doesn't mean anything in particular until it is read. Like every other text, Acts of Parliament are blueprints which readers - in this case, judges - must half decipher and half create by drawing on their linguistic competence and the other things they know. As trained experts in the law, judges know legal culture and tradition. But of course, that is not the only thing they know. And statutes are not written exclusively for judges. The cultural resources that individual readers bring to their reading depend on variable factors like education, religion, ethnicity, gender, class, and personal interests and experience. It is this variation in cultural resources that makes good faith disputes about meaning possible. As cultural variations grow more prevalent in a community, the official interpreters for that community must arguably become more self-conscious about the context they bring to a text. Reading contextually is not a choice; there's no other way to do it. On the other hand, acknowledging the limitations of a particular context, and where appropriate defending its strengths (or limitations), is a choice that judges can and should make.
My second criticism of the plain meaning rule is that it inappropriately directs attention to the problem of textual meaning and away from the many other issues that arise in disputes about interpretation. Although some disputes arise because of disagreement about meaning, often the disagreement is about
In a surprising number of cases, disputes purportedly decided by reference to the plain meaning of the text are in fact disputes about something other than meaning. In R. v. McIntosh, for example, the issue was not the meaning of s. 34(2), which was clear enough, but whether a mistake had been made in the drafting when this subsection was revised and re-enacted in 1954, and if so, whether it was the business of the court to correct it. To answer this question, the court needed to consider all reliable evidence of legislative intent. In R. v Multiform Manufacturing Co. Ltd.,(94) the Court had to determine whether a search and seizure provision in the Criminal Code could be relied on in an investigation by the Superintendent of Bankruptcy, in disregard of the search and seizure provision contained in the Bankruptcy Act. After pointing out that the Criminal Code provision applied in respect of "offences against this Act or any other Act of Parliament", Lamer C.J. wrote:
On a plain reading, s. 443 [of the Code] would thus apply to proceedings under any federal statute, regardless of whether or not the statute in question also contains search and seizure provisions. The use of the word "any" unambiguously shows that every single Act of Parliament could fall within the ambit of these provisions.(95)
What Lamer C.J. says is true, but the issue that required resolution was not the meaning of the words "any other Act of Parliament" but rather the relationship between the Criminal Code provision and the search and seizure provision in the Bankruptcy Act: did the latter oust the former? To answer this question, the Court needed to compare the two provisions and look for possible reasons to treat the provision in the Bankruptcy Act as an exclusive code.
In R. v. Thomson,(96) the court had to determine whether the Deputy Minister of a department was bound by the recommendations of a committee set up under the Canadian Security Intelligence Service Act to investigate a complaint about one of his decisions. The only relevant provision was the following:
52(2) On the completion of an investigation in relation to a complaint ..., the Review Committee shall provide the ... deputy head concerned and the complainant with a report containing any recommendations that the committee considers appropriate...
Cory J. wrote that the case turned on the meaning given to the word "recommendations": in this context, did "recommendations" mean binding directives or should the word be given its conventional meaning of non-binding suggestions? Not surprisingly, conventional meaning won. But in truth the real question here was not the meaning of "recommendations", but how the recommendations of the committee should be dealt with under the scheme. There was nothing in the legislation that expressly addressed this point; no provision telling the Deputy Minister what to do upon receiving a report with recommendations. In short, there was a gap in the legislative scheme. What needed to be decided was whether this was the sort of gap that courts could fill, and if so, whether the Court should fill it.
Knowing what the statute means is never enough because legislation is not self-applying. Once a court figures out the meaning, it must figure out the implications of the meaning by applying the statute to the facts. Even if a provision applies strictly as a matter of language, the court may decline to apply it as a matter of law - because the facts occurred outside the enacting jurisdiction, because they occurred before the coming into force of the provision, because legislation designed to prevent fraud cannot be used as an instrument of fraud or because for some other cogent reason applying the provision to these facts would be legally unacceptable. M.(K.) v. M.(H.)(97) is a good example of the Court exercising what is essentially an equitable jurisdiction and refusing to apply legislation when to do so would be unfair. Section 45 of Ontario's Limitations Act reads:
45.--(1) The following actions shall be commenced within and not after the times respectively hereinafter mentioned,
(j) an action for assault, battery, wounding or imprisonment, within four years after the cause of action arose;
Section 47 of the Act postpones the running of the limitation period for persons who suffer from one of four mentioned forms of legal disability, namely being a minor, being mentally defective, being mentally incompetent or being of unsound mind. In M(K) v. M(H), the Court added to the provision by holding that the period also does not run until the facts giving rise to the cause of action have been or should have been discovered by the plaintiff through the exercise of reasonable diligence. As La Forest J. explained, "the courts will not allow a limitation period to operate as an instrument of injustice."(98) This reasoning is cogent and it leads to a conclusion that many of us, though perhaps not all of us, find acceptable. The point to notice is that this outcome has nothing to do with the meaning of words and everything to do with the appropriate application of law.
One of the most striking things about the plain meaning rule is that although it purports to be a rule that governs the interpretation of statutes generally, it is applied only some of the time. It is applied regularly to fiscal legislation and often to penal legislation; but outside those contexts, it is readily shrugged off or ignored.
The Court's willingness to abandon the rule when it leads in the wrong direction is illustrated by its judgment in Rizzo v. Rizzo.(99) The issue in Rizzo was whether employees who lost their job because of their employer's bankruptcy were entitled to termination pay and severance pay under ss. 40 and 40a of Ontario's Employment Standards Act. The relevant parts provided:
40. (1) No employer shall terminate the employment of an employee ... unless the employer gives ... notice ....
(7) Where the employment of an employee is terminated contrary to this section,
(a) the employer shall pay termination pay ....
40a. (1a) Where
(a) fifty or more employees have their employment terminated by an employer in a period of six months or less and the terminations are caused by the permanent discontinuance of all or part of the business of the employer at an establishment
the employer shall pay severance pay .... [My emphasis.]
The Court of Appeal suggested that the plain meaning of these provisions makes termination by the employer a prerequisite of both termination and severance pay. In the Supreme Court of Canada, speaking for a bench that included Cory and Major JJ., Iacobucci J. wrote:
Consistent with the findings of the Court of Appeal, the plain meaning of the words of the provisions here in question appears to restrict the obligation to pay termination and severance pay to those employers who have actively terminated the employment of their employees. At first blush, bankruptcy does not fit comfortably into this interpretation. However, with respect this analysis is incomplete.
Elmer Driedger in Construction of Statutes (2nd ed. 1983) best encapsulates
the approach upon which I prefer to rely. He
recognizes that statutory interpretation cannot be founded on the wording of legislation
Although the Court of Appeal looked to the plain meaning of the specific provisions in question in the present case, with respect, I believe that the court did not pay sufficient attention to the scheme of the ESA, its object or the intention of the legislature; nor was the context of the words in issue appropriately regarded.(100)
Notice that Iacobucci J. does not claim that the language to be interpreted here is ambiguous and for this reason the plain meaning of the text cannot govern the outcome. He appears to concede that the text is plain. But in this case, for reasons that are never stated, he prefers an intentionalist approach. This casual disregard of a rule that is so insisted on in other contexts is hard to understand.
In some of its judgments the Court has tried to justify the special emphasis on plain meaning when interpreting fiscal legislation. The following passage from Principles of Canadian Income Tax Law, by P.W. Hogg and J.E. Magee, has been quoted and endorsed by the Court in several cases:
It would introduce intolerable uncertainty into the Income Tax Act if clear language in a detailed provision of the Act were to be qualified by unexpressed exceptions derived from a court's view of the object and purpose of the provision. [The Antosko case] is simply a recognition that "object and purpose" can play only a limited role in the interpretation of a statute that is as precise and detailed as the Income Tax Act. When a provision is couched in specific language that admits of no doubt or ambiguity in its application to the facts, then the provision must be applied regardless of its object and purpose.(101)
Hogg and Magee suggest that literal interpretation, uninformed by purposive analysis, should govern a court's approach to the Income Tax Act because that Act is drafted in a precise and detailed style. L'Heureux-Dubé J. suggests that literal interpretation is appropriate in tax cases because of the heavy reliance in the Act on technical commercial language. In Manulife v. Bank of Canada she writes:
The "modern contextual approach" is, in my view, the standard, normative approach to judicial interpretation, and one may exceptionally resort to the old "plain meaning" rule in appropriate circumstances. One example of the latter is statutory interpretation in the area of taxation, where the words and expressions used in legislative provisions quite often have a well-defined "plain meaning" within the business community.(102)
In other words, as the primary audience of many provisions in the Act, the business community should be able to rely on its own understandings of commercial language.
Certainly it is fair to suggest that provisions drafted in precise, concrete terms allow for fewer interpretive possibilities than provisions drafted in general or abstract language. It may also be fair to infer from a series of detailed provisions that the legislature intended to provide complete instructions for its audience, with minimal reliance on the discretion of interpreters. Finally, I would agree with L'Heureux-Dubé J. that the reasonable expectations of intended readers should be taken into account by judges. The problem with these analyses is the suggestion that they have special or exclusive application to the Income Tax Act. Some language in the Income Tax Act is detailed and precise, and some of it is technical, but these justifications for placing special emphasis on literal meaning hardly apply to all language in the Act. Many tax disputes turn on language that is as general or vague or ordinary as anything found in the Divorce Act or the Canada Labour Code. Furthermore the Income Tax Act has no monopoly on detailed provisions or precise technical language; these are found throughout the statute book and in principle should equally attract the plain meaning rule.
A further problem with these analyses is that they are inconsistent with the rule itself. According to the rule, the plain meaning of a text prevails if it has only one plausible meaning. In making this initial judgement, the type of legislation and the audience at which it is directed are supposed to be irrelevant. The determination of plain meaning is supposed to be based on linguistic competence alone. This is what makes the rule certain and the same for all.
Finally, to suggest that the Income Tax Act always or generally attracts a textualist approach contradicts the principle that "the interpretation of tax legislation should follow the ordinary rules of interpretation."(103) This principle was established by Estey J. in Stubart Investments Ltd v. The Queen(104) and it has since been reaffirmed by the Court.(105)
I have argued that it is arbitrary to insist on the plain meaning rule when interpreting tax or penal legislation, while rejecting it when interpreting other kinds -- or at least it appears to be arbitrary. However, there may be a consideration operating here, unarticulated and perhaps unconscious, that explains the practice of the Court. It seems that in these cases the plain meaning rule acts as a proxy for strict construction.
Historically the courts have embraced a policy of strictly construing both fiscal and penal legislation. This policy is not arbitrary, but is grounded in respect for private property, human freedom and the rule of law. While the state has the power to interfere with its subjects, to take money from them or put them in jail, the courts insist on certainty, clarity and fair notice in the exercise of this power. Strict construction of fiscal and penal legislation has been the main tool used by courts to protect these common law values. However, in some circumstances, the plain meaning rule can also be used for this purpose -- for example, when the language of the text is under-inclusive or the drafter has made a mistake that favours the subject. In such cases, the plain meaning rule justifies courts in giving primacy to the text over intention. It is at least arguable that this policy-based concern underlies the approach in R v. McIntosh and in some of the tax cases.(106) In fact, it appears that the recent reluctance of the Court to invoke the strict construction rule in these traditional areas is a major motive for the revival of the plain meaning rule.
In the criminal law context, the movement away from strict construction is evident in the Court's suggestion that strict construction should be relied on only as a rule of last resort.(107) In the tax cases, strict construction has purportedly been abolished.(108) While these attempts to diminish reliance on strict construction appear to defer to the legislature, in my view the appearance is deceptive. In fact, the courts continue to cherish the values underlying strict construction; these values remain an important part of the legal context in which legislative texts are interpreted. When they are not invoked openly and relied on directly, they are invoked and relied on in less direct and obvious ways. Thus, instead of asserting and justifying its mandate to protect private property from unexpected or unfair interference, the Court says that it is simply applying the text as written, as if it had no choice. This is not an improvement, in my view, and it is certainly not deference to the legislature.
Although in principle the plain meaning rule applies equally to all legislation, in practice the rule is invoked and insisted on in some contexts while readily ignored in others. Rather than serving as a general approach to interpretation, it is used as a technique to ensure appropriate outcomes in particular cases. The problem with using the plain meaning rule for this purpose is that, once again, it disguises what is really going on. The court purports to give primacy to the text when in fact it is giving primacy to common law values. And since the common law values are not acknowledged, there is no opportunity to justify them or qualify them by appealing to competing common law values.(109) The specific intention of Parliament, in so far as it might be discovered, gets lost in the shuffle.
In this part of my paper I have tried to show the flaws in the plain meaning rule. I would like to see it abolished from statutory interpretation. This does not mean that I think judges can ignore the wording of texts or can rewrite them at will. The ordinary meaning of a text sometimes seems compellingly clear and in such cases it carries significant weight. This is appropriate because rule of law is an important value of our legal system. However, even in such cases it is still important for judges to look to other sources of legislative intent as well as to relevant constitutional and supplementary law(110) to produce an appropriate outcome. These other factors are important because rule of law is not the only value in our legal system.
In the second edition of Construction of Statutes, Driedger wrote:
Today there is only one principle or approach, namely, the words of an Act are to be read in their entire context and in their grammatical and ordinary sense harmoniously with the scheme of the Act, the object of the Act, and the intention of Parliament.(111)
Since 1983, this so-called modern principle of interpretation has been cited and relied on in innumerable decisions of the Court. In recent years, however, it has become a source of considerable confusion.
In my view it is evident, from reading both the modern principle itself and the second edition generally, that Driedger rejected the plain meaning rule in favour of an intentionalist approach to interpretation. Like most literalists, he believed that the purpose of statutory interpretation is to discover and implement the intention of the enacting legislature. However, his conception of intention was much broader than theirs. In his chapter on the method of construction he wrote:
It may be convenient to regard "intention of Parliament" as composed of four elements, namely:
The expressed intention-the intention expressed by the enacted words;
The implied intention- the intention that may legitimately be implied from the enacted words;
The presumed intention- the intention that the courts will in the absence of an indication to the contrary impute to Parliament; and
The declared intention - the intention that Parliament itself has said may be or must be or must not be imputed to it.(112)
Proponents of the plain meaning rule recognize "the expressed intention"; for them this is the only evidence of legislative intent, at least when the words are clear. Driedger, however, in keeping with the intentionalist approach, relied not only on what is expressed but also on what is implied by the context and what is presumed by the common law.
For Driedger, "the fundamental principle [is] that the words of the Act must be read in their context".(113) This became the first instruction in his modern principle -- "the words of an Act are to be read in their entire context." Entire context includes the Act as a whole, and also the statute book as a whole, the body of law as a whole, and admissible external aids. It includes not only the "internal context" consisting of the language of the Act, but also the "external context" which Driedger examined under four headings: the social context, the intellectual context, the legal context and the language context.(114) We are far from literalism here. What Dreidger offered was a way to determine utterance meaning -- the speaker's intended meaning.
The second instruction in Driedger's modern principle is to read the words of the Act in their grammatical and ordinary sense harmoniously with the scheme and object of the Act and the intention of Parliament. The grammatical and ordinary sense of words was for Driedger their ordinary meaning, the meaning that would be understood by an ordinary reader immediately upon reading the text. This meaning, we are told, must be brought into harmony with the scheme and object of the Act -- the legislature's goals and the means chosen to achieve them. Notice that to find out the scheme and object of an Act and to bring the meaning into harmony with them, a purposive analysis is required. We are further told that the ordinary meaning must be brought into harmony with the intention of Parliament. We have seen that in Driedger's analysis this includes presumed intention, which in turn brings in the rest of the legal system -- international law, constitutional law, the general statute book, and the common law, all of which the legislature is presumed to comply with and not to want to change. At this point we are light-years from literalism.
Despite the intentionalist thrust of the modern principle, it has come to be identified with the plain meaning rule in recent judgments of the Court. This confusion can be traced in part to some remarks of Estey J. in the Stubart case:
Gradually, the role of the tax statute in the community changed, as we have seen, and the application of strict construction to it receded. Courts today apply to this statute the plain meaning rule, but in a substantive sense so that if a taxpayer is within the spirit of the charge, he may be held liable....
While not directing his observations exclusively to taxing statutes, the learned author of "Construction of Statutes", 2nd ed., ... put the modern rule succinctly:
Today there is only one principle or approach, namely, the words of an Act are to be read in their entire context and in their grammatical and ordinary sense harmoniously with the scheme of the Act, the object of the Act, and the intention of Parliament.(115)
As I read Estey J. in this passage and throughout his judgment in Stubart, he is saying that the old way of interpreting fiscal legislation, which relied on strict construction and the plain meaning rule, has been replaced by something new. Strict construction has "receded"; the plain meaning rule has given way to "plain meaning in a substantive sense"; and fiscal legislation is therefore to be interpreted, as any other legislation would be, purposefully and in total context.
This reading of Stubart is reflected in Notre Dame de Bon Secours.(116) However, beginning with Canada v. Antosko, a different understanding emerged. In the Antosko case, Iacabucci J. quotes the key passage from Stubart as authority for applying the plain meaning rule, which he understands in the following way:
While it is true that the courts must view discrete sections of the Income Tax Act in light of the other provisions of the Act and of the purpose ... such techniques cannot alter the result where the words of the statute are clear and plain....(117)
In other words, Iacobucci J. ignores "plain meaning in a substantive sense" and reverts to plain old plain meaning. He is followed down this path by Major J. in Friesen:
In interpreting sections of the Income Tax Act, the correct approach, as set out by Estey J. in Stubart Investments Ltd v. The Queen ... is to apply the plain meaning rule.(118)
Cory J. makes the same association in Alberta (Treasury Branches) v. Canada.(119) L'Heureux-Dubé J. makes it in Régie de permis d`alcool.(120)
This confounding of Driedger's intentionalist approach with the plain meaning rule is only partly due to the unclear writing of Estey J. in Stubart. Driedger himself invited the confusion at several points in his analysis. One of the confusing things he did was to redefine "literal meaning" so that instead of referring to a-contextual sentence meaning (as it always had before) it was now to refer to meaning-in-total-context . This redefinition permits the following conclusion:
It is clear today, the words of the Act are always to be read in the light of the object of the Act. Thus, the two approaches, Heydon's Case [purposive analysis] and Sussex Peerage [the plain meaning rule], have been combined into one... Today's doctrine is therefore still a doctrine of "literal" construction, but literal in total context and not, as formerly, literal in partial context only.(121)
One sees what Driedger meant, but this way of putting it is not helpful. If literal construction is now construction in total context, then perhaps it follows that the plain meaning rule is now the same as Driedger's modern principle.
The potential for confusion was increased when Driedger structured his five steps for the interpretation of legislation in a way that echoed the plain meaning rule. These steps are quoted with apparent approval by Lamer C.J. in McIntosh.(122) In slightly simplified form, they say:
Read the Act as a whole in its entire context so as to determine Parliament's intention, object and scheme.(123)
Read the words to be interpreted in their grammatical and ordinary sense in light of Parliament's intention, object and scheme.
If the words are clear and unambiguous and in harmony with the intention, object and scheme, and the general body of law, that is the end.
3b. If the words are obscure or ambiguous, adopt a meaning that the words are capable of bearing and that accords with the intention, object and scheme.
If the words are clear and unambiguous but not in harmony with the Act, other Acts or the general law, adopt a meaning that the words are capable of bearing and that produces harmony.
If the obscurity, ambiguity or disharmony cannot by resolved objectively by reference to the intention, object or scheme, then adopt the most reasonable meaning.
In my view, this set of instructions is too formal and elaborate to be helpful. But the point to notice is that for Driedger, interpretation was never to stop with first impression meaning. That meaning must be tested against and brought in harmony not only with the scheme and purpose of the Act but also with presumed intent and the body of general law. Disharmony no less than ambiguity requires resolution; and if a resolution cannot be found by "objective" means, interpreters must rely on their own reason. Although Driedger's steps echo the plain meaning rule in certain ways, the approach he described is a clear repudiation of that rule.
In his second edition, Driedger retained some of the terminology and rhetoric of the plain meaning rule, but the approach he adopted was intentionalist. Unlike the textualist who would prefer to focus exclusively on literal meaning, Driedger directed interpreters to wide ranging evidence of legislative intent, including extrinsic evidence and judge-made presumptions. In the revised third edition, the so-called modern principle is recast in the following terms:
There is only one rule in modern interpretation, namely, courts are obliged to determine the meaning of legislation in its total context, having regard to the purpose of the legislation, the consequences of proposed interpretations, the presumptions and special rules of interpretation, as well as admissible external aids. In other words, the courts must consider and take into account all relevant and admissible indicators of legislative meaning. After taking these into account, the court must then adopt an interpretation that is appropriate. An appropriate interpretation is one that can be justified in terms of (a) its plausibility, that is, its compliance with the legislative text; (b) its efficacy, that is, its promotion of the legislative purpose; and (c) its acceptability, that is, the outcome it produces is not unjust or unfair.(125)
Some courts have found significant differences between this formulation and Driedger's modern principle.(126) In my view, although rhetorically the formulations differ, substantially they differ in only minor respects.
One difference is that the second edition refers to the "modern principle" while the third edition refers to the "modern rule". Driedger called his approach a "principle" presumably to emphasize its generality and to set it apart from the "rules" of statutory interpretation which are described elsewhere in the book. However, the third edition suggests that the so-called rules of interpretation are actually not rules. They are principles which point interpreters in one direction or another and provide a justification for choosing that direction; but unlike rules, they are not binding.(127) The only genuine rule of statutory interpretation is that judges must exercise appropriate professional competence. They must resolve the interpretation dispute in accordance with the law, while remaining within the boundaries of their role as interpreters. Both the modern principle set out in the second edition and the modern rule set out in the third edition can be understood as attempts to give content to the notion of interpretive professionalism.
Both editions point out that interpretation begins with the text but emphasize that the text must be looked at in total context. Both acknowledge judicial reliance on legal values and principles and emphasize the importance of purposive analysis in every case. However, in the 2nd edition legal values and principles are analyzed under the rubric "intention of Parliament" which is defined to include presumed intent. In my view, given that it is judges who decide what Parliament is presumed to intend, this analysis is misleading. It blurs the real differences between specific intentions which a drafter is instructed to embody in particular legislation and the general values and principles which drafters must always take into account when drafting legislation. The third edition avoids this equivocal understanding of intention by treating the legal values and principles invoked by judges to resolve interpretation disputes as a separate category from intention.
Another difference between the modern principle and the modern rule lies in the way they deal with choice. Driedger's formula does not describe how judges make choices between competing possibilities because he envisaged harmony rather than disharmony. He imagined the various aspects of context all working together, pointing in the same direction. But what happens when the elements of his formula pull in different directions? If the ordinary and grammatical meaning of a text suggests one outcome, but the legislation's scheme and object suggest another, which prevails? And on what basis? The formula doesn't say.(128) The modern rule differs from the modern principle primarily in acknowledging the reality of choice. When relevant factors point in different directions, interpreters have no choice but to choose. They must weigh the competing factors and come down on one side or another -- this is where the judging comes in, and they must offer acceptable reasons for the outcome -- this is where the constraints come in.
It is evident that the modern rule set out the third edition reflects a pragmatic bias. However, it can be reconciled with an intentionalist approach to intepretation, just as Driedger's modern principle was, through presumption. If the principles and values appealed to in determining an appropriate outcome are presumed to be shared by the legislature, the outcome itself can be presumed to be intended by the legislature.
Despite assumptions to the contrary, Driedger's modern principle rejects the plain meaning rule in favour of an intentionalist theory of interpretation. For Driedger, simply reading the text was never enough. The third edition of Driedger reformulates his modern principle, but does not radically change it. It merely draws attention to the actual sources of presumed intent.
In the final part of this paper I will argue that pragmatism is a better theory of statutory interpretation than either textualism or intentionalism because it offers a more plausible account of the courts' jurisdiction in interpretation and a more accurate account of how outcomes are achieved. Pragmatism explains and justifies the current practice of Canadian courts without resort to linguistic or legal fictions. Textualism works as a theory of interpretation only if one accepts the false metaphor of language as a conduit. Intentionalism works only if one accepts the doctrine of presumed intent under which judicial values and choices are deemed to have been intended by the legislature. The chief virtue of pragmatism is that it recognizes the limitations of text and intention as determinants of outcomes in statutory interpretation cases and it reconciles the resulting judicial activism with democracy and the rule of law.
We saw in Part 2 of this paper that courts find it impossible to operate within the theoretical confines of textualism. Judicial attempts to focus on literal meaning are constantly undermined by express and implicit references to purpose, consequences and legal values. We have also seen that judges who insist on a textualist approach in some cases abandon it in others. The Rizzo case examined in Part 2 is hardly an isolated example. Even the strongest proponents of the plain meaning rule cast literal meaning aside when a more compelling consideration is present. In Friesen Major J. tells us that the correct approach to the interpretation of the Income Tax Act is to apply the plain meaning rule. However, in Schwartz he argues against the literal interpretation of s. 3(a) of that Act because it contradicts established case law:
Section 3(a) ostensibly permits taxation of income from any source. The argument of the Minister, which is supported by the literal wording of the section, is that "office, employment, business and property" are only examples of sources which may be taxed....
However, a literal adoption of this position would arguably constitute a dramatic departure from established tax jurisprudence.... Despite the inclusive language of ss. 3(a) and 56, ... Canadian courts have always recognized that monies which do not fall within the specifically enumerated sources are not subject to tax.(129)
In McIntosh and the C. P. case, Lamer C.J. insists that textualism is the only legitimate approach to interpretation.(130) But in Michaud v. Quebec,(131) he adopts an intentionalist approach, emphasizing the purpose of the provisions to be interpreted and the need to balance competing legal interests. One of the questions arising in Michaud was whether s. 187(1)(a) of the Criminal Code (dealing with the interception of private communications) permitted judges to open a sealed packet for the purpose of ruling on an application under the section. At the relevant time the provision read:
187 (1) All documents relating to an application [for authorization of surveillance] are
confidential and, with the exception
of the authorization, shall be placed in a packet and sealed by the judge to whom the application
is made immediately on
determination of the application, and that packet shall be kept in the custody of the court in a
place to which the public has
no access or in such other place as the judge may authorize and shall not be
(a) opened or the contents thereof removed except
(i) for the purpose of dealing with an application for renewal of the authorization, or
(ii) pursuant to an order of a judge of a superior court....
The provision does not say that a judge who is asked to open a packet may look at it in private and take its contents into account in deciding whether to make an order under subparagraph (ii). It says that packets may not be opened save in the circumstances mentioned. However, Lamer C.J. writes:
... a stark, literal reading of the provision would appear to suggest that the court must rule on such a motion while turning a blind eye to the contents of the packet.... In my view, the provision should be interpreted as permitting a judge to examine the contents of the packet in private for the restricted purpose of adjudicating a s. 187(1)(a)(ii) application. The confidentiality interests underlying the provision are simply not triggered when a competent judicial authority examines the contents of the packet in camera.(132)
Lamer C.J. here abandons the rigidities of textualism. To implement the purpose of the legislation in a fair and effective way, he is willing to add a third exception to the list set out in para. (a). In my view, this outcome is appropriate and is justified by the reason he gives. But it is important to notice that his approach here is not consistent with the plain meaning rule. Section 187(1)(a) of the Criminal Code is drafted in language that is as precise and detailed as legislative language gets; certainly it is no less plain than the provision interpreted in McIntosh. It even uses technical language. Yet it does not govern the outcome in this case.
In McIntosh, Lamer C.J. insists on the primacy of the text; the only question that interests him is the meaning of the words that Parliament has used. In Michaud, he relies on purposive analysis to fill a small gap in the legislative scheme. The focus is not on the meaning of the words used, but on the purpose of the scheme and its sensible operation. For the textualist the only measure of correct interpretation is the text; for the intentionalist, the only correct measure is intention. Neither theory allows interpreters to shift back and forth in this way. For the pragmatist, however, shifts in emphasis and focus are perfectly acceptable provided they are made for good reasons. The virtue of a pragmatic theory is that it requires judges to explain the legitimacy of such shifts by exposing their good reasons. Had Lamer C.J. adopted a pragmatist approach in Michaud, he could have justified his decision to add to the text by noting first, that the gap in question was a small one; second, that the provision was addressed to the judiciary rather than the public at large and was concerned with criminal procedure, an area in which the courts have inherent jurisdiction and special expertise; and third, that the scheme was well developed so that it was evident what bit had been overlooked and how the omission should be filled.
For intentionalists the task of the judge in statutory interpretation is to give effect to the intention of the legislature as inferred from reading the text in total context. Legislative intent is a powerful notion and should not be lightly dismissed. Although it has been harshly attacked by legal realists and others, at least some of the criticism is unfounded. Studies by linguists show that the process of inference-drawing relied on by courts to infer legislative intent is no different from the process relied on by the readers of any text. No reader ever has direct access to the writer's mind. Conclusions about writers' intentions are always speculations, based partly on the words of the text and partly on readers' knowledge. To the extent we can assume that this knowledge resembles the knowledge of the legislature in relevant respects, the speculations of readers are plausible and persuasive. Within limits, then, intentionalism offers an effective account of the court's task in interpretation.
However, there are two problem with the intentionalist account of interpretation. The first is that it imputes outcomes to the intention of the legislature not only in cases where this is plausible but equally in cases where the legislature never considered the issue in dispute. Often the intentions of the legislature, in so far as they can be known with reasonable certainty, do not suffice to resolve the problem before the court. The second problem with intentionalism is that it overlooks the supervisory role of the court. Sometimes the court declines to apply a provision to facts because to do so would lead to unacceptable results. In cases like these, the outcome is not based on intention but depends on other factors. In pretending otherwise, the intentionalist denies responsibility for her own choices and loses credibility with her audience.
Because legislation is drafted in the form of general rules that will apply to facts which the legislature has not even attempted to imagine, the plausible inferences of judges concerning legislative intent often stop short of providing the answer to the question facing the court. In such cases, judges must make it up. They must become secondary law-makers, drawing on the resources of the existing law, on evolving social and cultural values and on their own sense of justice and right reason.
The necessity for secondary law-making can be illustrated by looking at the recent judgment of the Court in Opetchesaht Indian Band v. Canada.(133) In that case the Court had to determine the validity of a permit issued to the British Columbia Power Commission. This permit gave the Commission a right to erect and maintain power transmission lines on an Indian reserve "for such period of time as the said right-of-way is required for the purpose of an electric power transmission line." The issue was whether the permit was validly issued under s, 28(2) of the Indian Act which provides:
28(2) The Minister may by permit authorize any person for a period not exceeding one year, or with the consent of the council of the band for any longer period, to occupy or use a reserve or to reside or otherwise exercise rights on a reserve.
Subsection 28(2) is an exception to the general prohibition against alienation found in s. 37 of the Act.
37. Except where this Act otherwise provides, lands in a reserve shall not be sold, alienated or otherwise disposed of until they have been surrendered to her Majesty ...
The surrender process referred to here is an elaborate one designed to ensure that the interests of the band are appropriately protected.
A textualist would resolve the issue of validity by asking what is the meaning of "period" in s. 28(2). If "period" is understood to mean "a length of time consisting of a stipulated period of years", then the right-of -way is unauthorized and invalid; however, if "period" means "a length of time bounded by the happening of an ascertainable event", then the right-of-way is fine. An intentionalist would resolve the issue by asking whether the legislature intended long term permits of the sort at issue here to be dealt with under s.28(2) or under s. 37. If the former, the permit is valid; if the latter, it is not.
Notice that neither question can be answered with assurance. It is impossible to say that "period" in this co-text plainly means one of these things and not the other. And absent discussion of this very issue in Parliament, or other relevant evidence, who can say what Parliament intended? Clearly the legislature intended some grants to be exempt from the onerous surrender provisions designed to protect the band. Clearly it intended to include grants relating to rights excisable on a reserve for some period longer than one year. But there is no indication that the legislature specifically considered grants of rights-of-way that could last indefinitely. There is no basis either in the text or its context for drawing a relevant inference. Thus, neither text nor intention compels a particular outcome. Each of these factors must be taken into account and given as much weight as it can bear. But since neither is determinant in the circumstances, the Court must ultimately rely on other considerations. And it does.
McLachlin J. emphasizes the historical relationship between the Crown and Aboriginal peoples which underlies the protection of Indian interests found in the Act(134) as well as the principle of interpreting legislation in favour of Aboriginal peoples.(135) She attaches no weight to the interests of the utility or to the public interest in having efficient, low cost power.(136) She concludes that a permit of such long-term consequence to the Aboriginal community is outside the scope of the section. Major J. also recognizes that Aboriginal interests must be protected, but for him the autonomy of the band in making decisions about its is a more compelling concern.(137) He also notes, though without comment, that the permit at issue in the case is one of over a thousand similar arrangements made between bands and utilities across the country.(138) He concludes that a temporary permit granting a limited right in land is within the scope of the section. The different conclusions turn on the different ways of assessing the competing interests and policies.
The supervisory role of the courts is openly acknowledged when they are called on to dispose of jurisdictional challenges to the validity of enactments or decisions. It is less evident, but equally present and important, when courts determine the meaning of legislation and test the consequences of applying legislation to particular facts. If applying a provision to facts would produce unacceptable outcomes -- outcomes that are unconstitutional, irrational, incoherent, unjust or unfair -- the courts may choose to not apply it.
Although this last statement may sound controversial, there are many non-controversial ways in which courts avoid unacceptable outcomes. The presumption of rationality, for example, plays an essential and pervasive role in the formation of first impression meaning. Although it generally operates at a subconscious level, this presumption is a major determinant of meaning. If you read the sentence "Andrew found a mole under the hedge", you are likely to think first of the underground mammal, rather than a person engaged in espionage or the skin growth. If the sentence appears in the context of a story about pests, you are unlikely to notice the ambiguity of the word. At a more conscious level, avoiding absurdity is a standard reason for resolving perceived ambiguity one way rather than another. Take the following sentence: "No one shall fire a weapon at another person except for a police officer". Although grammatically it might be more natural to attach the exception for police officers to "person" rather than "no one", to avoid the unthinkable most interpreters would resolve the ambiguity the other way.
In addition to the general presumption of rationality, the courts rely on traditional and evolving common law presumptions, many of which are non-application rules. For example, it is presumed that the legislature wants to comply with the Charter and other limits on its jurisdiction, that it observes international law, that it does not concern itself with trifles, that it enacts workable schemes, and that it respects individual freedom, private property rights and the rule of law. Provisions are presumed not to apply to facts outside the territory of the enacting jurisdiction, to facts that occurred before the provision came into force or to agents of the Crown. Relying on these presumptions, interpreters are free to narrow the scope of general words or discover implicit exceptions to broad, unqualified provisions.
In R. v. Hinchey,(139) for example, both the majority and minority judgments relied on common law doctrine to narrow the broad language of s. 121(1)(c) of the Criminal Code. This provision made it an offence for a government employee to accept, directly or indirectly, "a commission, reward, advantage or benefit of any kind" from " a person who has dealings with the government" unless the written consent of the employer has been obtained. The concern of the Court was the very wide sweep of the section, which potentially might criminalize innocent or trivial behaviour. L'Heureux-Dubé J., speaking for the majority, wrote:
... that the judiciary should not declare innocent conduct criminal is a principle of statutory construction which decrees that Parliament does not intend through the criminal law to trap trivial, non-criminal conduct. As Gonthier J. expressed in Ontario v. Canadian Pacific Ltd........because the legislature is presumed not to have intended to attach penal consequences to trivial or minimal violations of a provision, the absurdity principle allows for the narrowing of the scope of the provision.(140)
Similarly Cory J. wrote that the boundaries of the section must be defined "to ensure that it did not encompass conduct of the accused which no reasonable member of the community would regard as blameworthy."(141) Thus even though a person employed by the government receives something from a person who is attempting to negotiate a supply contract with the government, if the two turn out to be neighbours and the thing received is a cup of coffee, the provision does not apply.(142)
In Hinchey and elsewhere these common law doctrines are labeled presumptions of intent so that giving effect to them becomes by definition giving effect to the legislature's intent. This works on a formal level, but there are costs to this approach. First, treating common law doctrines as a form of legislative intention mutes judicial responsibility and makes it less likely that these presumptions will be critically examined. It makes them harder to challenge and harder to change. Second, it tends to empty the notion of "legislative intent" of real significance or force. It is difficult to take the notion seriously if in fact it is judges who make it up.
I have argued that the chief virtue of pragmatism is that it recognizes the limitations of the traditional theories. This has two positive effects. First, it requires judges to acknowledge and justify the other factors that influence their judgments. Second, it requires judges to address the crucial problem of weight. If the only acceptable determinant of outcome is text or legislative intention, then in principle the problem of weight does not arise; one or the other factor governs. This makes justifying outcomes easy, but unpersuasive. And ironically, it devalues the very factor that is said to be determinant. If all outcomes are said to flow from the text, or alternatively from legislative intention, than the genuinely constraining aspects of the text or intention are not singled out and assessed. When the approach is formalistic, deemed constraints have the same force as the real thing. Pragmatism is more persuasive because it asks judges to attribute only as much weight to the text as it can bear having regard to considerations like relative clarity, and only as much weight to legislative intention as it can bear having regard to the cogency of the inferences that can be drawn from reading the text in context and from relevant extrinsic evidence. Pragmatism expects judges to also rely on other, judicially noticed factors but only to the extent such reliance is warranted have regard to the relevance and importance of these other factors. The clearer the text, the more compelling the evidence of specific legislative intent, the harder it will be to justify departure from those constraints and the more compelling the other factors will have to be.
Another virtue of pragmatism is that it offers a better way of characterizing the separation of powers between legislatures and courts. Under traditional versions of the separation doctrine, the legislature makes the law and the court's job is to apply that law to particular facts. Under the pragmatic approach, the legislature makes statutes and the court's job is to resolve disputes in accordance with the law. On this approach statutes are the primary source, but not the only source of law. Judges apply the relevant statute but they also apply constitutional law and what is sometimes called supplemental law. In a common law jurisdiction supplemental law consists of the rest of the statute book, the common law and the evolving legal tradition which draws on current social and political values as well as those of the past. In a civil law jurisdiction supplemental law consists of the basic Codes, the rest of the statute book, legal doctrine and the evolving legal tradition.(143) Pragmatism alone of the three theories we have looked at accommodates the crucial role played by constitutional and supplemental law in statutory interpretation.
The final virtue of pragmatism is that it reconciles the reality of judicial choice with the imperatives of democracy. As the Supreme Court of Canada has recently pointed out, there is more to constitutional democracy than government by elected representatives of the people:
... the evolution of our constitutional arrangements has been characterized by adherence to the rule of law, respect for democratic institutions, the accommodation of minorities, insistence that governments adhere to constitutional conduct and a desire for continuity and stability.(144)
The courts play a major role in ensuring adherence to these values; this role is what gives them their legitimacy. Thus, even if it were possible, it would be wrong for courts to apply the directives of the legislature mechanically and blindly to particular facts. Courts must act as mediators to ensure that disputes are resolved in ways that respect all the values of constitutional democracy. This point is made most eloquently by the Court in Reference re Secession of Quebec:
[A] system of government cannot survive through adherence to the law alone. A political system must also possess legitimacy, and in our political culture, that requires an interaction between the rule of law and the democratic principle. The system must be capable of reflecting the aspirations of the people. But there is more. Our law's claim to legitimacy also rests on an appeal to moral values, many of which are imbedded in our constitutional structure. It would be a grave mistake to equate legitimacy with the "sovereign will" or majority rule alone, to the exclusion of other constitutional values.(145)
In my view, that is precisely the problem with both textualism and intentionalism: they equate legitimacy with the majority rule expressed in statute law to the exclusion of other sources of legitimacy, including the appeal to fundamental political, social and moral values.
The pragmatic account of statutory interpretation is often resisted because it is thought to give too much discretion to unelected judges, contrary to both democratic principle and the rule of law. But in truth pragmatism does not "give" judges additional discretion; it merely acknowledges the discretion they have, and must have, to resolve interpretation disputes. The only real question is whether we are content with formalism or want to know the "real" reasons for decision. In a homogenous community, where the same cultural values are shared and not much challenged, formalism works. In such communities outcomes really do seem to spring from the very words of the text; the intent of the legislature seems self-evident. But in the community of Canada at the end of the 20th century, the inadequacies of formalism are apparent. The Court's judgments are widely read and reported to the public at large. They are assimilated into legal culture and presented to the larger community by a profession and media that increasingly reflect the divergent realities of modern life. The Supreme Court of Canada cannot simply point to the text when the text in fact means different things to different people. Nor can it simply invoke the intention of the legislature when the governing factors are clearly rooted in judge-made law. To be credible these days, the Court must acknowledge its choices and draw on all the sources of its legitimacy to persuade its diverse audience, consisting of the parties, the profession and the public, that its choices are appropriate.
1. See 2747-3174 Québec Inc. v Québec (Régie des permis d'alcool),  3 S.C.R. 919, at pp. 995-6.
2. My analysis draws on William Eskridge and Philip Frickey, "Statutory Interpretation as Practical Reasoning " (1990) 42 Stanford Law Rev. 321 , which surveys the judicial theories of statutory interpretation in U.S. case law under the headings "intentionalism", "purposivism" "textualism" and "practical reasoning ". However, in Canadian there is no systematic distinction between intentionalism and purposivism.
3. The plain meaning rule is sometimes referred to as the "literal meaning rule" or the "literal rule": see Pierre-André Côté, The Interpretation of Legislation in Canada, 2nd ed. (Cowansville, Quebec: Les éditions Yvon Blais, 1992) at pp. 237-255.
4. Members of the court who regularly invoke the plain meaning rule are Lamer C.J., Cory J., Iacobucci J., and Major J. Gonthier J., L'Heureux-Dubé J. and MacLachlin J. regularly reject it. The position of Batarache J. and Binnie J. has not yet been established.
5. See, for example, Paul Campos, "That Obscure Object of Desire: Hermeneutics and the Autonomous Legal Text" (1993) 77 Minn. Law Rev. 1065; Barbara Child, "What Does 'Plain Meaning' Mean These Days?" (1992) 3 Scribes Journal of Legal Writing 1; Clark Cunnighma, Judith Levi, Georgia Green and Jeffry Kaplan, "Plain Meaning and Hard Cases" (1994) 103 Yale Law J. 1561; Judith Levi et al, "Northwester University/Washington University Law and Linguistics Conference" (1995) 73 Washinton Univ. Law Q. 769; D'Amato "Counterintuitive Consequences of 'Plain Meaning'" (1991) 33 Arizona Law Rev.; William Eskridge, "The New Textualism" (1990) 37 U.C.L.A. Law Rev. 621; Daniel Farber, "The Inevitability of Practical Reason: Statutes, Formalism and the Rule of Law" (1992), 45 Vanderbilt Law Rev. 533; Fish, Stanley, "The Anti-formalist Road" in Doing What Comes Naturally - Change, Rhetoric and the Practice of Theory in Literary and Legal Studies (Durham: Duke Univ Press, 1989); Catherine Fisk, "The Last Article About the Language of ERISA Preemption? A Case Study of the Failure of Textualism" (1996) 33 Harvard J. on Legislation 35; Allen Hutchinson, "A Postmodern's Hart: Taking Rules Sceptically" (1995) 58 Modern Law Rev. 788; Bradley Karkkainen "' Plain Meaning:' Justice Scalia's Jurisprudence of Strict Statutory Construction" (1994) 17 Harvard J. of Law & Public Policy 401; John Manning "Textualism as a Nondelegation Doctrine" (1997) 97 Columbia Law Rev. 673; Jerry Mashaw "Textualism, Constitutionalism and the Interpretation of Federal Statutes" (1991) 32 William & Mary Law Rev. 827; Paul Perell "Plain Meaning for Judges, Scholars and Practioners (1998) 20 Advocates Quarterly 24; Richard Pierce "The Supreme Court's New Hypertextualism: An Invitation to Cacophony and Incoherence in the Administrative State" (1995) 95 Columbia Law Rev 749; John Polich "The Ambiguity of Plain Meaning: Smith v. United States and the New Textualism" (1994) 68 So. California Law Rev 259; William Popkin "An Internal Critique of Justice Scalia's Theory of Statutory Interpretation" (1992) 76 Minnesota Law Rev. 1133 Raymond Randolph, "Dictionaries, Plain Meaning, and Context in Statutory Interpretation" (1994) 17 Harvard J. of Law & Public Policy 71; Frederick Schauer "Statutory Construction and the Coordinating Function of Plain Meaning" (1990),  Supreme Court Review 231; Lawrence Solan "Learning Our Limits: The Decline of Textualism in Statutory Cases" (1997) 97 Wisconsin Law Rev. 235; Muriel Spence "The Sleeping Giant: Textualism as Power Struggle" (1994), 67 So California Law Rev. 585; nicolas Zeppos "Justice Scalia's Textualism: The 'New' Legal Process" (1991) 12 Cardoza Law Rev. 1597.
6. For an excellent introduction to these studies, see Geoffrey Underwood and Vivienne Batt, Reading and Understanding: an Introduction to the Psychology of Reading (Oxford: Blackwell, 1996).
7. Elmer Driedger, Construction of Statutes, 2nd ed. (Toronto: Butterworths Ltd., 1983), at p. 87.
8.  2 S.C.R. 1028, at pp. 1049-50. My emphasis.
9.  1 S.C.R. 686, at p. 705.
10. See R. v. Creighton, [ 1993] 3 S.C.R. 3 at ; R. v. McIntosh, supra note 9, at pp. 703-5; Ontario v CP Ltd, supra note 8, at pp. 1051, 1053, and 1055.
11. For a striking example, see R. v. Paul  1 S.C.R. 641, at pp. 662-4.
12. McLachlin J. is an intentionalist; when L'Heureux-Dubé J. and Gonthier J. are not being pragmatic, they are intentionalists too.
13. Supra note 9, at pp. 712-3.
14. Supra note 1, at pp. 1001-5.
15. Ibid., at p. 1001.
16. P. Michell, "Just do It! Eskridge's Critical Pragmatic Theory of Statutory Interpretation" (1996) 41 McGill L.J. 713 at p. 731, quoted in Regie d'alcool, ibid. at p. 1010.
18. Ibid., at p. 1001.
19.  2 S.C.R. 606.
20. Supra note 8, at p. 1070; see also pp. 1083-4.
21. Supra note 19, at p. 641. My emphasis.
22. This term is fully defined and discussed infra at p.
23.  1 S.C.R. 254.
24. Ibid., at p 296.
25. In the Oxford English Dictionary "loss" is defined as "the fact of losing" and "losing" is defined "to fail to obtain (e.g. a prize)". Cf . R. v. Dawson,  3 S.C.R. 783, where for the majority of the Court the argument turned on the meaning of "deprive" in "intent to deprive [a parent] of the possession" of a child. It was argued that a person cannot be "deprived" of something that he or she does not have. However, this argument was rejected. L'Heureux-Dubé J. who wrote, at p. 796, that "to deprive a person of something means, among other things, to keep that person from that which he or she would otherwise have: Oxford English Dictionary, supra, vol. IV, at p. 490."
26. For another recent example of a case in which the choice of text could affect the outcome, see R. v. Audet,  2 S.C.R. 171, at pp. 192 & ff, where the Court identifies the text as consisting of "trust" as opposed to "a position of ... trust".
27. See Jacob L Mey, Pragmatics: An Introduction (Oxford: Blackwell, 1993) at p. 184: "Usually, one defines the co-texts of a (single or multiple) sentence as that portion of the text which (more or less) immediately surrounds it. (Unfortunately, there are no agreed limits as to what "immediately" is supposed to mean here.)"
28. For discussion of the role of working memory in reading, see supra note 6, at pp. 24-8.
29.  3 S.C.R. 72.
30. Ibid., at p. 80. See also Ontario v. C.P. Ltd, supra note 8, where Lamer C.J. implies that plain meaning of the text to be interpreted is properly established without co-text. Referring to the DeSousa case, he writes, at p. 1053: "the majority interpreted "unlawful act" as requiring objective foreseeability of body harm... -an interpretation that itself clearly departs from the 'plain meaning' of the word 'unlawful act' standing alone." My emphasis.
31. Supra note 8, at p. 1050. My emphasis.
32. Ibid., p. 1054. My emphasis.
33. In fact, in determining the plain meaning of "use" in clause 13(1)(a), Lamer CJ. does not consider the statute as a whole, but refers merely to the rest of the subsection.
34. Supra note 8, at p.1054.
35. Supra note 29, at p. 80.
36. There is certainly no justification for this assumption. Linguists do not know how lexical meanings are stored in from the brain. None of the models currently under consideration resemble dictionaries.
37. Heriete de Swart, Introduction to Natural Language Semantics (Stanford: Centre for the Study of Language and Information, 1998), p. 8.
38. Ibid., p. 9.
39. This sense of ordinary meaning is adopted in the third edition of Driedger.: see Ruth Sullivan, Driedger on the Construction of Statutes, 3rd (Toronto: Butterworths Ltd., 1994) at pp.
40. For illustration and discussion of such a survey, see Clark Cunningham et al, "Plain Meaning and Hard Cases" (1994) 103 Yale L.J. 1561, which was cited by the United States Supreme Court in United States v. Granderson (1994) 114 S.Ct. 1259 at 1267; United States v. Staples (1994) 114 S.Ct. 1793 at 1806; and Director, Office of Workers' Compensation Programs v. Greenwich Collieries (1994) 114 S. Ct. 2251 at 2255. See also Clark Cunningham and Charles Fillmore, "Using Common Sense: A Linguistic Perspective on Judicial Interpretation of 'Use A Firearm'" (1995) 73 Washington Univ. Law Q. 1159.
41. For a self-conscious illustration of such speculation, see Perrier Group of Canada Inc. v. Canada,  1 F.C. 586, at p. 597, where Linden J.A. concluded that Perrier water is a "beverage" within the meaning of Schedule III of the Excise Tax Act partly on the basis of the following analysis: "If a server in a Canadian restaurant asked a customer which 'beverage' to bring and the customer responded, 'Perrier, please', would the server be surprised that the customer thought that Perrier was a beverage? I think not. Would the server respond to the customer saying, 'Perrier is a water, and I shall bring it, but do you want a beverage as well?' I think not. In our common speech, most Canadians, in my view, would include water, especially sparkling water, within the meaning of beverage, despite the many dictionary definitions excluding it."
42. Some theorists do take notice of this type of meaning. The best discussion I have come across is by Jorge Gracia in A Theory of Textuality: The Logic and Epistemology (Albany, New York: State Univ. of N.Y. Press, 1995 at pp. 164 and ff. Gracia writes at p. 164:
Interpretations whose main or only purpose is to produce understandings of the meanings of texts ... may be distinguished from a second sort. That second sort are interpretations whose primary aim is not to produce such understandings, even in cases when such understandings are necessary for the fulfillment of the primary aim.
The aim of the second sort of interpretation, which Gracia calls non-textual interpretation, depends on its cultural function. For such interpretation, understanding the meaning of the text is only the beginning; the interpreter must than explore the relation of the text to other things. Examples of non-textual interpretation mentioned by Gracia include historical, legal and literary interpretation.
43. See, for example, supra note 8, at pp. 1050-1, 1054, and 1055, where Lamer C.J. uses the terms "literal meaning" and "plain meaning" interchangeably.
44. See  1 S.C.R. 963, at p. 976, where Cory J. refers to the absence of "any doubt as to the meaning of the legislation nor any ambiguity in its application to the facts". My emphasis.
45. See  3 S.C.R. 415, at p. 438, where L'Heureux-Dubé J. refers to language in legislation having "a well-defined 'plain meaning' within the business community."
46. See, for example, R.v. Audet, supra note 26, at pp. 193-4, where LaForest J. wrote:
In the absence of statutory definitions, the process of interpretation must begin with a consideration of the ordinary meaning of the words used by Parliament. Le Grand Robert de la langue francaise (2nd ed. 1986) defines the French word "autorité" as [trans] right to command .... The Oxford English Dictionary (2nd ed. 1989) suggests similar definitions for the English word "authority".... As can be seen from these definitions, the ordinary meaning of the word "authority" or "autorité" does not permit so restrictive an interpretation [as suggested by counsel for the defence].
47. See, for example, Schwartz v. Canada, supra note 23, at p. 296, where La Forest J. wrote:
The statutory requirement that one must be "in the service" of another person to be characterized as an "employee" excludes ... any notion of prospective employment when the phrase is given its ordinary meaning. An employee is "in the service" of his or her employer from the moment he or she becomes under obligation to provide services...; however, employment does not necessarily begin from the moment the contract is entered into....During oral argument, counsel admitted that an ordinary person would find that Mr. Schwartz was not an employee of Dynacare when the contract was cancelled.
48. Friessen in (1)
49. Supra note 9, at pp. 687-8.
50.  2 S.C.R. 312, at p. 327. My emphasis.
51.  3 S.C.R. 103, at p. .
52.  4 S.C.R. 707 at p. 719. My emphasis.
53.  2 S.C.R. 119, at p. 152.
54. Ibid., at p. 154.
55.  3 S.C.R. 3, at p.
56. Supra note 44, at p. 976.
57. See, for example, Royal Bank of Canada v. Sparrow Electric,  1 S.C.R. 411, at p.442.
58. Supra note 9, at p. 704.
59.  3 S.C.R. 727.
60. Ibid., at p. 764.
61. Ibid., at pp. 764-5.
62. In having a preferred outcome and then trying to justify it, judges are not being inappropriately subjective or impartial. Judicial preferences, like other professional judgements, are intuitive responses grounded in legal knowledge and training and professional experience.(63)
63. For discussion of literature on how experts make decisions and its relevance to judicial decision-making, see Daniel Farber, "The Inevitability of Practical Reason: Statutes, Formalism, and the Rule of Law" (1992) 45 Vanderbilt Law Rev. 533 at pp. 554 & ff. - " "
64. Supra note 9, at p. 696.
66. Ibid., at p. 697.
68. See, for example, Jorge Gracia, A Theory of Textuality: the Logic and Espistemology (Albany, New York: Statute University of New York Press, 1995), p. 29: "Generally the shorter and less complex a text is the more different meanings it may have depending on context."
69. Notice that this criticism cannot be answered by pointing out that courts take judicial notice of meaning. Taking judicial notice of the meaning of words must be distinguished from taking judicial notice of the meaning of a particular text and both are different from taking judicial notice that the meaning of a particular text is plain. The latter does not meet the test.
70. Cory J is occasionally an exception . For example, in Alberta (Treasury Branches) v. Canada, supra note 44, at pp. 976-7, he says "the very history of this case with the clear differences of opinion expressed as between the trial judges and the Court of appeal of Alberta indicates that for able and experienced legal minds, neither the meaning of the legislation nor its application to the facts is clear."
71. Supra note 8, at p. 1055.
72. Ibid., at p. 1081.
73. Georgia Green, Pragmatics and Natural Language Understanding, 2nd ed. ( Mahwah, New Jersey: Lawrence Erlbaum Ass., 1996) p.10.
74. See Michael J. Reddy, "The Conduit Metaphor - A Case of Frame Conflict in Our Language About Language" in Andrew Ortney (ed.), Metaphor and Thought (Cambridge: Cambridge University Press, 1979) at pp. 284-324.
75. Green p. 10
76. PA Prather and David Swinney "Lexical Processing and Ambiguity Resolution: AN Autonomous Process in an Interactive Box" in Lexical Ambiguity Resolutio: Perspectives from Psycholinguistics, neuropsychology & Artificial Intelligence ed by Steven Small, Garrison Cottrell and Michael Tenenhaus (San Mateo, Calf: Morgan Kaufmann Publishers: 1988) at pp 289-310 at 290.
77. Grag B Simpson and Curt Burgess, "Implications of Lexical Ambiguity Resolution for Word Recognition and Comprehension" in Lexical Ambiguity Resolutio: Perspectives from Psycholinguistics, neuropsychology & Artificial Intelligence ed by Steven Small, Garrison Cottrell and Michael Tenenhaus (San Mateo, Calf: Morgan Kaufmann Publishers: 1988)
271-288 at 276-7. See also Green who writes at pp. 49-50: "...it is possible to show that virtually every noun is polysemous (indeed, virtually every verb, adjective, and preposition as well), and possibly infinitely so....Furthermore, in context, just about any noun can be used to refer to just about any sort of thing."
78. Jorge Gracia, A Theory of Textuality: the Logic and Espistemology (Albany, New York: Statute University of New York Press, 1995), p. 28.
79. Frank Smith, Understanding Reading: A Psycholinguistic Analysis of Reading and Learning to Read, 5th ed. (Hillsdale, New Jersey: Lawrence Erlbaum Ass., 1994) at p. 31.
80. Schemes (also called "frames") are general patterns, regularities or stereotypical situations that occur in our experience, are stored in memory, and form the basis for expectations and predictions and appropriate behaviour. Scenarios (or scripts) are standard sequences of events; in effect, they are schemes that unfold through time. An example of a scheme is knowledge about the lay-out of a fast food restaurant and the type of food available; an example of a scenario is how one goes about ordering a hamburger in such a place. For discussion of these, and related concepts, see Brown and Yule, "Representing background knowledge" pp.236-256. See also Smith, especially chapters 1-3 and 9.
81. Genre refers to distinct types of writing. Plays, poems, personal letters, business memoranda, government reports, and bathroom graffiti are examples of different genres. Each has its own conventions of objects, content, structure, layout, language, register, cohesion, and the like. Frank Smith writes at p. 44: "To be able to read a text, we must be able to anticipate the conventions that the writer will use. This understanding of the appropriate conventions, with prior knowledge relevant to the subject matter, is what readers contribute. To be comprehensible the writer must anticipate and respect the conventions that the reader will predict."
82. François Rastier, Meaning and Textuality (trans. F. collins and P. Perron)(Toronto: University of Toronto Press, 1997) at pp. 6. See also p. 30.
83. Ruth M Kempson, Semantic Theory (Cambridge: Cambridge University Press, 1977) at p. 100.
84. Gillian Brown and George Yule, Discourse Analysis (Cambridge: Cambridge University Press, 1983) at p. 11.
85. Frank Smith, p. 12
86. De Swart, p.10. My italics.
87. Rastier, p. 5. My italics.
88. Gracia p.29.
89. Smith p. 167
90. Smith p. 3
91. Smith p. 161
92. Stephen Crain and Mark Steedman, "On not being led up the garden path: the use of context by the psychological syntax processor" in David Dowty et al, Natural Language Parsing (Cambridge: Cambridge University Press, 1985) 320-358 at p. 338.
93. Green, p. 59
94.  2 S.C.R. 624.
95. Ibid., at p. 631.
96. (1992) 89 D.L.R. (4th) 218 - find SCR cite
97.  3 S.C.R. 6.
98. Ibid., at pp. 58-9.
99.  1 S.C.R. 26.
100. Ibid., at pp. 40-1.
101. See Friesen v Canada,  3 S.C.R. 103 at 112-4; Alberta (Treasury Branches) v Canada, supra note 44, at p. 976 ; Royal Bank of Canada v Sparrow Electric Corp.,  1 S.C.R. 411, at p.
102. Supra note 45, at p. 438.
103. Supra note 54, at p.20.
104.  1 S.C.R. 536, at pp. 576-8.
105. See, for example, supra note 23, at p. 295.
106. Lamer C.J. draws attention to the connection himself in R. v. McIntosh supra note 9, at p. , when he wrote, in justifying his refusal to accept absurdity as grounds for reading in words which do not appear on the face of the legislation: "The Criminal Code is ... qualitatively different from most other legislative enactments because of its direct and potentially profound impact on the personal liberty of citizens. The special nature of the Criminal Code requires an interpretive approach which is sensitive to liberty issues." See also Major J. in Alberta (Treasury Branches) v Canada, supra note 44, at p.
107. See R. v. Hasselwander  2 S.C.R. 398, at p. 413. See also Reference re Criminal Code Sections 193 & 195.1(1)(c),  4 W.W.R. 481, at p. 512.
108. See Stubart Investments Ltd. v. The Queen, supra note 103, at p. 578 ; (Québec) Communauté Urbaine v. Corporation Notre-Dame de Bon-Secours and authorities cited, supra note 54, at pp. 14-7.
109. John Mark Keyes has drawn my attention to R. v. Chartrand,  S.C.R. , as a good example of a case in which the Court self-consciously considers and prefers competing values, in this case the protection of children from abduction.
110. These terms are explained infra at
111. Driedger, 2nd ed. P. 87
113. Ibid. p. 107.
114. Ibid. pp. 107-8.
115. Supra note 103, at p. 578.
116. Supra note 54, at pp. 15-7, 20.
117.  2 S.C.R. 312, at pp. 326-7.
118.  3 S.C.R. 103, at p. 113.
119. See supra note 44, at pp. 975-6.
120. See supra note 1, at p. 996.
121. Supra note 7, at p. 83.
122. Supra note 9, at p. 698-9.
123. Notice that Driedger does not say to read the entire Act to figure out the meaning of the words, but rather to figure out the purpose and scheme.
124. Supra note 7, at p.105. My emphasis.
125. Ruth Sullivan, Driedger on the Construction of Statutes, 3rd ed. (Toronto: Butterworths Ltd., 1994), p.
126. See, for example, Haida Nation v. British Columbia (Minister of Forests) (1997) 153 D.L. R. (4th) 1 (B.C.C.A.), at pp. 8 & ff.. Esson J.A. writes, at p. 8: "The differences [between Driedger's modern principle and the modern rule of the 3rd edition] are not merely in style or form. The significant difference is in the substance". In fact, the 3rd edition turns Driedger's principle on its head, he suggests, "by reducing the language of the enactment to a matter of minor importance." He correctly points out, at p. 10, that the literalists on the Supreme Court of Canada continue to cite the 2nd edition of Driedger and make no mention of the 3rd edition.
127. Ibid at
128. Driedger does acknowledge the possibility of disharmony in his five steps, set out in the text supra at note 121. Step 4 says that if the words point one way but other features point another way, adopt a meaning that produces harmony and is linguistically plausible. But there may not be such a meaning. Step 5 acknowledges the possibility that harmony and linguistic plausibility may not be achievable in every case. In "hard" cases the court is to adopt the most reasonable meaning. These two steps are not adequately reflected in the "modern principle" that is quoted so widely by the courts.
129. Supra note 23, at p. 300.
130. See infra at pp.
131.  3 S.C.R. 3.
132. Ibid., at pp. 30-1.
133.  2 S.C.R. 119
134. Ibid., at pp. 154 & ff.
135. Ibid., at p. 153.
136. Ibid., at p. 163.
137. Ibid., at p. 145.
138. Ibid., at p. 130.
139.  3 S.C.R. 1128.
140. Ibid., at p. 1152.
141. Ibid., at p. 1180.
142. Cory J. also relied on the doctrine of mens rea which is now grounded in the Charter. He pointed out, ibid. at p. 1181, that it offers an alternative method of limiting the scope of the section so that "only morally blameworthy activity comes within its purview."
143. See R.A. Macdonald "Harmonizing the Concepts and Vocabulary of Federal and Provincial Law: The Unique Situation of Quebec Civil Law" in The Harmonization of Federal Legislation with Quebec Civil Law and Canadian Bijuralism: Collection of Studies (Canada: Department of Justice, 1997) pp. 31-67.
144. Reference re Secession of Quebec,  SCJ 61 para 48
145. Reference re Secession of Quebec,  SCJ 61 para 67. My emphasis.
Please feel free to contact M. Anne Vespry
if you can suggest other resources.