Legal Drafting Website: Statutory Interpretation Resources

The Plain Meaning Rule and Other Ways to Cheat at Statutory Interpretation

by Ruth Sullivan
Professor of Law
University of Ottawa


Anyone who spends time studying language and interpretation soon comes to appreciate the complexity and uncertainty inherent in any effort to interpret a text . While in some contexts (the study of poetry, for example) complexity and uncertainty are valued, when it comes to disputes about the interpretation of legislative texts, these features are a cause of dismay -- and a certain amount of dissembling. It is this last response to complexity and uncertainty that I focus on in this presentation. I am interested in the ways superior courts, which are the official interpreters of legislative texts, attempt to disguise the complexity and uncertainty of interpretation with doctrines like the plain meaning rule and fidelity to legislative intent. Although these doctrines do not fully or accurately reflect what judges must do to resolve interpretation disputes, they generate a reassuring rhetoric. If we can't have simple messages and certain outcomes, reassuring rhetoric is the next best thing.


There are two key assumptions relied on by courts to explain and justify their work in statutory interpretation. One is the assumption that meaning inheres in legislative texts and that at least some of the time meaning is "plain" -- that is, clear and certain, not susceptible of doubt. This assumption is the necessary basis for the plain meaning rule. The other assumption is that legislatures have intentions when they enact legislation and these intentions are knowable by courts when called on to interpret legislation. This assumption is the necessary basis for the doctrine of fidelity to legislative intent.

Although these assumptions are repeatedly challenged by academics like myself, the courts are not inclined to give them up, for actually they are quite useful. The assumptions tell us that interpretation is rooted in something definite -- the text -- which has been fixed once and for all by the legislature. They thus support a view of statutory interpretation that fits the current mythology of western democracy. In a democracy such as ours, persons and their activities are governed not by the whim of rulers but by the rule of law. We know that the rule of law is legitimate because law is made by the elected representatives of the people, at the end of a democratic process. We know that the rule of law is fair and efficient because law is issued in the form of texts which are fixed and stable, published in advance and applied equally to everyone under the supervision of the courts. Finally, we know that courts are well placed to supervise the application of law because of their impartiality and their interpretive expertise.

It is obvious to anyone who reads the newspapers or otherwise follows public affairs that rule of law and the related ideals of democracy, fairness, equality and efficiency are highly prized by Canadians. These are the virtues in which government tries to dress itself and there is little incentive for those who are part of government to challenge the mythology or to suggest that those in power are not wearing clothes. Like the other branches of government, the judiciary is reluctant to draw attention to possible discrepancies between the rule of law ideal and its own practice in interpreting and applying legislation. Furthermore, there are dangers in too much iconoclasm. If law or its enforcement is perceived to be illegitimate or unfair, or if compliance with law is perceived to be pointless or impossible, there will be little incentive to obey the law. And despite our large police forces and crowded jails, the rule of law in western democracies is fundamentally self-enforcing. This is why the rhetoric of law is important. Any attack on the assumptions and doctrines which sustain our ideals must therefore proceed with care.

In this presentation I argue that the plain meaning rule and the doctrine of fidelity to legislative intent are misleading in certain ways and force the courts to rely on a number of "tricks" which disguise the important role played by discretion in statutory interpretation. I identify and describe these tricks and give some examples of how they are done. I complain that the courts' reliance on these tricks is a form of "cheating". However, mindful of the dangers of too much iconoclasm, I end by suggesting that there is really no need for the courts to cheat. With fairly minor revisions to their basic assumptions, it would be possible for courts to acknowledge the role of discretion in statutory interpretation without putting the virtues of democracy and rule of law in jeopardy.


The plain meaning says that if the meaning of a legislative text is plain, the court may not interpret it but must simply apply it as written. The court may to resort to the rules and techniques of interpretation only if the text is ambiguous.

This rule presupposes that there is an important difference between the first-impression meaning of a text and post-interpretation meaning. First impression meaning is meaning that spontaneously comes to mind when a person reads a text relying on nothing but the text and her own linguistic competence. Post-interpretation meaning is meaning constructed by a person through interpretation, by relying on factors other than the text itself -- factors like the imagined purpose of a text, or its possible consequences, or extrinsic aids like legislative history. According to the plain meaning rule, when a person sets out to resolve a dispute about the correct interpretation of a legal text, the first thing she must do is read the text and form an impression of its meaning based on reading alone. She must then judge whether this meaning is plain. A text has a plain meaning if a competent reader would judge, on the basis of reading alone, that her first impression meaning is the only meaning the text can plausibly bear. A text is ambiguous if a competent reader could plausibly read it in more than one way.

When a legislative text has a plain meaning, the courts are prohibited from interpreting it; they are bound by the words of the text. Given this fundamental rule, it is perhaps surprising that even when the meaning is plain, the courts may look at non-textual evidence of legislative intent (like purpose or consequences or extrinsic aids) and may rely on these other factors to support the plain meaning. What they may not do is rely on these factors to vary or contradict the plain meaning of the text. This approach to extra-textual evidence offers some evident advantages to proponents of the plain meaning rule: heads they win, tails they still win. I call this Plain Meaning Rule (PMR) Trick #7. The other PMR tricks are examined below.

The key feature of the plain meaning rule, and the feature that sets it apart from other approaches to statutory interpretation, is its claim that the court's first responsibility is to give effect to the apparent meaning of a legislative text whenever this meaning is plain. If there is a conflict between what the text appears to say and what the legislature seems to have intended, the text wins. Intention governs only when the text itself is ambiguous.


The great advantage of the plain meaning rule is that, in theory at least, it creates a zone of certainty -- an interpretation-free zone, in effect. It tells the public that if the text is plain, it means what it says and it is safe to rely on it. The courts won`t come along and trick you at the last minute by importing unsuspected qualifications or implications into the text, even if these qualifications or implications were probably intended by the legislature. This emphasis on text at the expense of intention ensures that the law is certain and that the public has fair notice, both of which are prerequisites for effective law.

A second advantage of the theory is that it supports formal equality. If the courts are bound by plain meaning, and plain meaning is the same for everyone, it follows that a rule whose meaning is plain expresses the same law for everyone and will be applied in the same way, to the same effect to everyone. Plain meaning thus creates a level playing field on which we all have equal opportunity to score.

A third advantage of the plain meaning rule is that it can be used as an apparently neutral proxy for strict construction. It is no coincidence that the plain meaning rule is applied most persistently and enthusiastically to fiscal legislation, with penal legislation running a close second. The courts have always been reluctant to assist the other branches of government in limiting the freedom of subjects or taking away their property. Historically in interpretation these judicial values have been given bite primarily through the doctrine of strict construction. However, in many instances a court can achieve the same result by relying on the plain meaning rule without having to invoke the common law values. By claiming that the text is plain and does not require interpretation, a court purports to base the outcome in a case on linguistic competence alone. The appeal to judge-made law and policy is replaced by an exercise in technical expertise.


Although the plain meaning rule is often evoked, it is not accepted by everyone. For many judges, the touchstone of interpretation is the intention of the legislature. The legislature may reveal its intentions directly, for example by explaining them in a preamble or a purpose statement. Some would argue that Ministerial statements addressed to the legislature during the passage of a bill also provide direct evidence of legislative intent. Generally speaking, however, legislative intent is discovered indirectly through inference: having regard to xxx, a person who said aaa in context yyy when speaking to audience zzz must have meant bbb. To draw inferences of this sort, even competent speakers must rely on a wide range of contextual factors, both textual and extra-textual.

If it appears that a text has a clear precise meaning, an intentionalist judge would normally attach significant weight to that factor, but she would still be obliged to consider other factors such as the legislative evolution of the text or its real world consequences. If this extra-textual evidence of legislative intent turned out to be sufficiently compelling, the intentionalist would have to accept it. What is distinct and important about intentionalism is that it gives top priority not to the apparent meaning of the text, but to the meaning it was intended to have. The intentionalist does her best to carry out the real intentions of the legislature.

Intentionalism treats legislation as if it were an ordinary speech act, that is, an act of communication occurring at a particular place and time in which a speaker attempts to evoke particular meanings in the mind of an audience. It assumes that speaker and audience are operating within a context that is more or less shared or is at least accessible to both. This context consists of shared linguistic, social and cultural conventions and shared historical experience. Intentionalism further assumes that the speaker is aware of the context in which the audience will carry out its interpretation and ensures accurate communication by anticipating the inferences the audience will draw when reading different combinations of words in that context. Although there is always the possibility of a misunderstanding or mistake, the speaker relies on the audience to interpret cooperatively.

A cooperative interpreter makes a good faith effort to reconstruct the actual intentions of the speaker, sorting out contradictions when necessary, correcting the speaker's mistakes, looking for the implications implicit in her choice of words. If I say "it's hot in here", I may intend simply to report on the temperature in my immediate surroundings; or I may intend to communicate my discomfort; or I may "really" be asking the other person in the room to open the window or turn on the fan. It's even possible that I wanted you to turn on the air conditioner and my use of "fan" was just a slip of the tongue. A cooperative interpreter attempts to piece together the speaker's actual meaning from all the available evidence .

Judges who accept the doctrine of fidelity to legislative intent are driven to accept a number of corollary doctrines including the original meaning rule, the doctrine of presumed intent and the distinction between interpretation and amendment. It is easy to see why an intentionalist approach to interpretation requires something like the original meaning rule. If a legislative text is designed to communicate a particular meaning, one that was actually intended by the legislature on the occasion of enactment, that meaning must be fixed once and for all at the moment of enactment; it cannot subsequently change. It is also easy to see why the doctrine of presumed intent is important to the intentionalist approach. This doctrine is needed to establish the legislature's connection to the context in which judicial interpretation occurs. A court can infer the meaning that was actually intended by the legislature only if its inferences are grounded in assumptions, values and conventions that were actually shared by the legislature at the time of enactment. By presuming the legislature intended to comply with whatever policies, values and conventions are relied on by courts from time to time, this connection is assured. Finally, it is easy to see why an intentionalist draws a sharp distinction between correcting or reading down a text so that it conforms to the actual intention of the legislature and reading in or filling gaps so as to supplement or change that intention. Legislative intent is a fact waiting to be discovered; and the job of the court is to discover it, not make it up.


The main function of intentionalism, and its chief advantage, is that it explains the role of courts in interpretation without challenging our popular conceptions of democracy. Most of us accept that for law to be legitimate it must be democratic, and for law to be democratic it must be made by the elected representatives of the people. If this is so, then the only legitimate source of law is the legislature; judge-made law is undemocratic and therefore illegitimate. While this reasoning has the potential to cause some serious problems, under the doctrine of fidelity to legislative intent, the question of judicial law-making is answered before it can arise. Judges do not make law; they only apply it. Being impartial and unbiased, they use their legal interpretation skills to discover and give effect to the intention of the legislature. They have neither the need nor the mandate to engage in political choice.

A second advantage of the intentionalist doctrine is that it supports the positivist view of law, which is a major bulwark of certainty. Positivism tells us that law is one thing and not-law is something else entirely and we can easily tell them apart by applying a simple rule of recognition. What the legislature enacts is law. Legal values, preferred policies, judicial assumptions -- these things are not law and therefore play no role in the governance of people's lives.


By suggesting that outcomes are dictated by a fixed text or a fixed legislative intent, the plain meaning rule and the doctrine of fidelity to legislative intent both promise certainty. The rules are embodied in black and white; and within the interpretation-free zone at least, they are the same for everyone. What you see is what you get.

Even when the meaning is not plain and interpretation is required, the judges who carry out this interpretation are constrained by the fixed intention of the legislature. Upon enactment, the content of the law is fixed once and for all; through interpretation the judges discover that content -- not change it, not create it -- and ensure that it applies equally to all.

At least this is how it goes in rule of law heaven. What are things like here on earth?


Here on earth it is easy to appreciate the rhetorical value of the plain meaning rule and the doctrine of fidelity to legislative intent. It is also easy to appreciate the importance of rhetoric in a system that relies primarily on hegemony rather than force to induce compliance. Nonetheless, despite the benefits they bestow, I object to these doctrines because in my view they force the courts to "cheat". Since this is strong language, I should take a moment to explain what I mean by cheating.

When a student cheats on a paper or exam, she takes someone else's answer and tries to pass it off as her own; she tries to claim the credit for someone else's work. Cheating in this sense is a combination of stealing and misrepresentation. I don't suggest that judges cheat in this way. In fact, what judges do is rather the reverse. They take their own answer-- the outcome in a statutory interpretation dispute -- and try to pass it off as someone else's. They try to deny responsibility for their work. Cheating in this sense is not a form of stealing, but it is a form of misrepresentation.

When judges invoke the plain meaning rule, they deny responsibility for the outcome by saying that the text made them do it. "The text is clear, its meaning is plain, I have no choice. If you don't like the outcome, tell the drafter who wrote the text or the government officials who told her what to write or the legislature which enacted the text into law." When judges invoke the doctrine of legislative intent, they transfer responsibility to the legislature. "I don't make the laws, I only apply them. I am a legal technician, a mere servant of my legislative master. In democracies like Quebec or Canada, the legislatures call the shots."

These doctrines are objectionable because they discourage judges from probing and discussing the real basis for outcomes in statutory interpretation disputes. They encourage judges to deny or misrepresent the choices they are obliged to make and to avoid responsibility for outcomes. In my utopia, judges acknowledge their responsibility and attempt to explain and justify their choices.

These complaints about "cheating" are hardly new. Complaints of this sort have been made over and over again, off and on, for most of this century. For the most part they are ignored by the courts. The question that interests me in this presentation is how the courts get away with it. Usually when someone is caught cheating, they are obliged to take a different approach. But the courts go right on invoking the same doctrines and using the same rhetoric year after year, more or less ignoring academics like myself. I think I know why they do it. They do it for the reassuring rhetoric. By relying on these doctrines, they provide credible resolutions to interpretation disputes without upsetting any applecarts.

The question I want to consider in greater detail in is how they do it. I want to draw attention to the techniques used in statutory interpretation to make choice look like choicelessness, to make discretion look like constraint, and to make judge-made law look like the intention of the legislature. I will start with the plain meaning rule, which in my view is sly and pernicious and deserving of much criticism. I will then look at some of the problematic aspects of intentionalism, in particular the original meaning rule, the doctrine of presumed intent, and the equivocal distinction between drafting mistakes and legislative oversights. My presentation will end with a call for a better rhetoric, one that doesn't force judges to cheat.


PMR Trick #1. Artful Text Selection

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 of statutes. The first step in applying the plain meaning rule is to identify the text-to-be-interpreted. This consists of the words whose meaning has been put at issue by an attempt to apply legislation to particular facts.

In R. v. McCraw,(1) for example, the facts were that an accused wrote letters to young women which said (using cruder and more aggressive language), "I'm going to have sex with you even if I have to rape you". The legislation was s. 264.1 of the Criminal Code which said:

Everyone commits an offence who, in any manner, knowingly utters ... a threat
(a) to cause death or serious bodily harm to any person;
(b) to burn, destroy or damage real or personal property; or
(c) to kill, poison or injure an animal or bird that is the property of any person.

In applying this language to these facts, everyone agreed that the young woman was a "person" and that the letter contained a "threat" that was "knowingly uttered". Everyone also agreed that the threat was not a threat of "death" or of interference with property in any of the forms mentioned in the text. The only thing the parties were prepared to argue about was whether the letter was a threat of "serious bodily harm". These then were the words in issue, the text-to-be-interpreted.

If you pay close attention, you'll notice that sometimes identifying the text-to-be-interpreted involves choosing between alternative texts which favour different outcomes. When this happens, the outcome of the dispute cannot really be blamed on the meaning of the text, for it actually depends (at least in part) on the initial choice of text.

To illustrate the point, we can look at the analysis of LaForest J. in R. v. Audet.(2) This case concerned a 22 year old man who, shortly after his teaching contract ended, had sexual contact with a 14 year old girl whom he had taught the previous year in eighth grade. At the time of the sexual encounter the accused was aware that his contract would be renewed the following year and that the girl might become his student again. He was accused of sexual exploitation contrary to s. 153 of the Criminal Code:

153. Every person who is in a position of trust or authority towards a young person ... and who for a sexual purpose touches ... the body of the young person ... is guilty of an indictable offence....

At trial the accused was acquitted on the ground that he did not occupy a position of trust or authority at the time of the sexual contact. Both the trial judge and the Court of Appeal for New Brunswick focussed on the words "position of"; however, in the Supreme Court of Canada LaForest J. focussed on the word "authority". To determine the meaning of this text, he looked up the words "autorité" and "authority" in French and English dictionaries and reached the conclusion that term was broad enough to include any exercise of power, not just those stemming from a formal relationship. He 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.... [T]he meaning of the term ["authority"] must not be restricted to cases in which the relationship of authority stems from a role of the accused but must extend to any relationship in which the accused actually exercises such a power. As can be seen from these definitions, the ordinary meaning of the word "authority" or "autorité" does not permit so restrictive an interpretation.(3)

Although La Forest J. implies that the fate of the accused was sealed by the meaning of the word chosen by Parliament, in fact he had a choice. He could have focussed on the words "in a position of". He could have pointed out that Parliament put these words in for a reason, then pointed to the usual reference of the words, namely a formally defined or officially created status.

I am not suggesting that the outcome preferred by a majority of the court was wrong. My point is that the outcome was chosen, not dictated by the plain meaning of "authority".

If the outcome of a dispute can differ depending on which words receive attention and the judge does not have to explain the basis for her choice, or even acknowledge that she's made one, the plain meaning rule's promise of certainty is rather empty in my view. This does not mean that outcomes are the result of whim or accident or that texts are meaningless and do not constrain at all. But it does suggest that there may be more to reading a text than proponents of the plain meaning rule are prepared to acknowledge.

PMR Trick #2. Elastic Co-Text

"Co-text" is a term from linguistics that refers to the text immediately surrounding the text-to-be-interpreted. More precisely, it is the amount of surrounding text that a reader needs to hold in his or her short term memory in order to make sense of the words. If you come across the word "shower" in isolation, you can't know what kind of shower is referred to. You can't even know whether the word is a verb or a noun. You need some co-text to form a definite impression of its meaning. In reading the following statements, notice how you use the co-text around "shower" to draw inferences about its meaning, and notice the different sorts of knowledge you rely on in drawing those inferences..

  1. a shower Since there is an article in front of the word, it must be a noun.

  2. He needed a shower. It's probably not a meteor shower. Although one could devise a set of circumstances in which a meteor shower could be needed, those circumstances would be highly unusual. Conversely, there are other types of shower that are commonly needed.

  3. The farmer needed a shower. It's probably not a wedding or baby shower since the practice of giving showers of this sort is normally associated with brides or new parents, not farmers.
  4. The farmer needed a shower after all that planting. It's either a rain shower (needed to encourage the seeds to germinate) or the kind you take indoors (needed to wash away dirt and sweat). But which?
  5. The farmer needed a hot shower after all that planting. He was cold and tired. Aha! It's the kind you take indoors. Outdoor showers are usually not "hot" and usually not relied on to warm a person up or soothe tired muscles. Conversely, indoor showers often are hot and used for these purposes.

For me, the meaning of "shower" was uncertain until I got to statement (5). But at that point I became pretty certain that "shower" means the kind taken indoors in a bathtub or a shower stall. My certainty was the result of having eliminated the other possibilities as implausible given what I know about language, farmers, planting, and who needs or receives different kinds of showers under different circumstances. Each word in the co-text triggered an additional set of associations, assumptions, experience and the like.

For purposes of the plain meaning rule, the co-text is the amount of surrounding text that a reader takes into account in deciding (1) the first impression meaning of the text-to-be-interpreted and (2) whether it is plain. If you pay close attention, you'll notice that in judgments that rely on the plain meaning rule, the co-text is marvellously elastic. Sometimes it includes a single section. Sometimes it expands to take in the entire statute. But more often it shrinks to nothing at all, when the court insists that the text-to-be-interpreted must be looked at in isolation.

The McCraw case offers an example of a text with no co-text. The issue in the case, you'll recall, was whether a threat of rape was an offence under s. 264.1. of the Criminal Code. The text-to-be-interpreted was "serious bodily harm". Cory J. determined the meaning of this text by adopting the definition of "bodily harm" in the Code and looking up "serious" in the dictionary. He then refused to look at anything else - including the other words of the section. Counsel for the accused argued that because "serious bodily harm" was coupled with "death" in the relevant provision, the sort of harm the legislature had in mind must be serious indeed. Cory J. refused even to entertain this argument. He 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.(4)

In this passage Cory J. implies that to determine the plain meaning of "serious bodily harm", the text must be read in isolation, ignoring the rest of the sentence in which it appears or at least not letting the rest of the sentence affect the reader's understanding.

In R. v. McIntosh Lamer C.J. took the same approach as Cory J. The text to be interpreted consisted of the introductory words of s. 34(2) of the Criminal Code. Lamer C.J. conceded that if you look at that text in light of ss. 34 and 35 together, it is contradictory and confusing. But if you look at it "in isolation" its meaning is plain.(5) However, contrast this with what Lamer C.J. said in Ontario v. C. P. Ltd., a case decided the very same year:

... 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.(6)


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).(7)

Does the size of the co-text matter? Is it really worth bothering about a technicality of this sort? In my view, the answer is yes. It is through the subtle manipulation of such technicalities that the courts manage to do one thing while appearing to do another. The size of the co-text often affects its meaning, and it is even more likely to affect judgements about ambiguity. Let's return to the provision considered in McCraw:

264.1 Everyone commits an offence who, in any manner, knowingly utters ... a threat
(a) to cause death or serious bodily harm to any person;
(b) to burn, destroy or damage real or personal property; or
(c) to kill, poison or injure an animal or bird that is the property of any person.

Suppose the question is whether a threat to rape is a threat to cause serious bodily harm and you are allowed to look at the words "serious bodily harm" but no others. How will you answer? Can you answer with any degree of assurance? Now suppose you are allowed to take in a certain amount of co-text, consisting of rest of paragraph (a) - a threat to cause death or serious bodily harm. Does that affect your answer? Would you feel more or less inclined to convict? Now let's expand the co-text a bit more, to include paragraphs (b) and (c)? Does this make any difference? In my view, it does. In my view, any heightened sense of seriousness created by the reference to "death" in para. (a) is thoroughly flattened by the reference to such relatively unserious matters as damaging property and injuring pets in the remainder of the provision.

In the C.P. case, Lamer C.J. concedes that co-text matters. He says that when he considers the word "use" in isolation it's ambiguous, but when he considers it in the context of the statute as a whole, it's not.(8) Given the way co-text works, this is hardly surprising. The more co-text you have, the easier it is to narrow down the possible meanings of a text. The thing to notice here is the way the courts can affect the meaning of a text, or affect how plain it appears to be, by shrinking or expanding the co-text. The plain meaning rule cannot provide certainty if the co-text is manipulable in this way.

PMR Trick #3 The Shifting Meaning Game

Anyone who reads statutory interpretation cases soon notices that the courts have a dozen and more different ways of referring to meaning. There's ordinary meaning, literal meaning, common sense meaning, ordinary and grammatical sense, natural sense, and so on. 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. This ever-shifting terminology is a shell game. When I say plain meaning, I might mean any of the following:

  1. "dictionary" meaning. Dictionary meaning is a-contextual word meaning.
  2. "literal" or "facial" meaning. Literal meaning is a-contextual sentence meaning, i.e., meaning that can be derived from the words alone and their arrangement in a sentence having regard to the rules of language alone and disregarding extra-textual factors like purpose or consequences.
  3. "intended" meaning (i.e., the meaning intended by the legislature). Linguists call this utterance meaning. It differs from literal meaning in that it depends on context, which is defined for this purpose as including all the knowledge that is stored in the minds of speakers and audiences. Drafters rely on their knowledge (of everything- not just language and not just law, but geography, economics, social convention, cause and effect -- everything ) to predict how the readers of legislation will interpret their words. These readers rely on their own encyclopaedic knowledge to reconstruct what the drafter (and legislature) must have intended when she wrote (and it enacted) those words.
  4. "audience-based" meaning. This is the meaning that is in fact understood by the audience for which the legislation was written - 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 usually called "ordinary" meaning. When the audience is a specialized sub-group, audience-based meaning is so-called "technical" meaning. When the audience is the specialized sub-group consisting of lawyers and judges, audience-based meaning is "legal" meaning.
  5. "applied" meaning. Applied meaning is the meaning of a text in relation to particular facts. If your job is to solve an interpretation dispute, then strictly speaking, this is the only meaning that matters.

When judges use the expression "plain meaning", what kind of meaning are they talking about? Dictionary meaning? Literal meaning? Applied meaning? If you pay close attention, you'll discover that there's no consistency in their usage. "Plain meaning" can refer to any of the types of meaning defined above -- or to none in particular. Often the "plain" or "ordinary" meaning of a text is equated with the dictionary definition of its words, or at least the portion of the definition that strikes the court as appropriate. Given that dictionaries set out the entire range of meaning that words are capable of bearing, it is ironic that dictionaries are the tool of choice when a court wants to put responsibility for an outcome on the plainness of the text. The court's dictionary research may or may not be supplemented by textual or contextual or purposive analysis. It may or may not be supplemented by speculation about the legislation's intended audience and how it might understand the words.

In practice the courts have no fixed understanding of what they mean by "meaning" and no standard approach to establishing it in particular contexts. They are free to use the dictionary or ignore it; to consider the audience for which the legislation was written or ignore it; to appeal to context or purpose or stick to the literal meaning, all as they see fit. These options create choice. But because the same vague terminology is indiscriminately applied to all the different possibilities, the choices being made are not apparent.

Along with artful text selection and elastic co-text, the vagueness of the judicial concept of meaning ensures that courts have significant discretion when it comes to determining first impression meaning but they are not obliged to acknowledge this discretion and indicate the factors governing its exercise.

PMR Trick #4 It Must Be Plain To You If It's Plain To Me

The distinction between plain and ambiguous, which lies at the heart of the plain meaning rule, requires judges to determine whether a text can plausibly bear more than one meaning. The judicial notion of plausibility is as elusive as the judicial concept of meaning. The courts make no attempt to give content to the notion. It appears that plausibility is a purely linguistic judgement. Or is there a legal dimension here as well? What factors are relevant to judgements about plausibility? Does perspective matter? Are there degrees of plausibility? If so, how are the degrees measured and classified? These questions are never addressed. The courts simply pronounce on the matter: this proposed meaning is plausible; that one is not.

So what happens when Judge A says "the meaning is this and this is the only plausible meaning" and Judge B says "pardon me, but you're mistaken, the meaning is plainly that." If we assume (as we must) that both judges are competent speakers of language and that neither is operating in bad faith, then it is impossible to say purely as a matter of language, that only one of them is right. It would seem to follow that in such a case the text cannot be plain, it must be ambiguous - in which case neither of them is right. In practice, however, proponents of the plain meaning rule generally ignore such empirical evidence of ambiguity. They prefer to rely on their personal linguistic intuitions.

A striking example of this disregard for empirical evidence can be found in the judgment of Lamer C.J.. in Ontario v. C.P. Ltd. In that case both Gonthier J. for the majority and Lamer C.J. for the dissent defined 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 considered the meaning of this text in light of the same co-text, the subsection as a whole and both carried out a bit of textual analysis. Lamer C.J. concluded:

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".(9)

After reading the same text in the same co-text, Gonthier J. concluded

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.(10)

In other words, Gonthier J. thought that the meaning of the expression "for any use that can be made of [the natural environment]" is "for any non-trivial use ...".

Unlike Gonthier J., Lamer C.J. concluded that the meaning he identified was the only meaning the text could plausibly bear. Notice what this says about Gonthier J.'s analysis. It says that the meaning identified by Gonthier J. (and accepted by a majority of the Court) is implausible; it implies that Gonthier J. and the judges who concurred with him lacked linguistic competence. One must presume that Lamer C.J. did not mean to imply this, that he would readily acknowledge in his colleagues a linguistic competence comparable to his own. So what then does he mean? By declaring his preferred meaning plain and invoking the plain meaning rule, Lamer C.J. purports to settle the case on linguistic grounds . The preferred meaning of Gonthier J. is ostensibly rejected on linguistic grounds. Yet if challenged on linguistic grounds, Lamer C.J. would surely back away.

This purported reliance on linguistic plausibility is perhaps the most objectionable aspect of the plain meaning rule. By declaring the meaning plain on the basis of her own intuition, even though that intuition is not shared by others, a judge is able to dismiss the arguments that work against her without ever having to deal with them. Competing interpretations are summarily dismissed as being linguistically implausible; and competing arguments need not be considered because plain meaning governs, even in the face of evidence that the legislature intended something else.

PMR Trick # 5 The Inherent Meaning Illusion

The plain meaning rule is based on the fundamental proposition that once a text is written, it "contains" a meaning that does not change, but remains stable regardless of the context in which the text is read and regardless of the different knowledge and expectations readers bring to the text. Recent work in linguistics has discredited this view.(11) Texts do not "contain" meanings; rather they invite readers to infer meanings from the words chosen, the arrangement of the words, the circumstance of utterance and other aspects of context. The text of legislation may be fixed once and for all, but the context cannot be.

Context consists of the linguistic and social conventions which the reader has internalized and the broad range of cultural assumptions absorbed through family, school, religion, work and leisure activities like reading and watching television. It includes everything the reader knows or thinks she knows, everything that is "stored" in her mind at the time of reading. While much context is shared with other readers, context also varies from one time or place to the next, and one reader to the next. These differences in context explain how different readers can look at the same text and infer different meanings. As the linguists Yule and Brown explain:

However objective the notion of "text" may appear ... 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).(12)

If the linguists are right (and judges are not really in a position to challenge them), the assumption that texts contain a fixed meaning turns out to be false; it is merely a convenient fiction. If the meaning of a text is not in the text, but must be constructed through inference that depends in part on open-ended contexts, the distinction between reading and interpretation cannot be sustained.

This is not good news for proponents of the plain meaning rule, for the rule's promise of certainty is rooted in and depends entirely on the fixed meaning assumption. If we are to have certainty, the meaning of a text must be the same for everyone who reads it. If meaning is not fixed by the text, but can vary depending on context, then it is does not have to be the same for everyone. The certainty promised by the plain meaning rule turns out to be an illusion.

Notice that getting rid of the notion of a fixed inherent meaning does not destroy meaning or preclude communication. It is still possible for readers to experience particular meanings as clear or plain. And if these readers are working within similar contexts, they will probably experience the same particular meanings as being plain. This is what happens in successful communication. Getting rid of the notion of a fixed inherent meaning does not preclude successful communication, but it should require interpreters to acknowledge the complexity and uncertainty that are unavoidable in attempts to communicate.

PMR Trick # 6 Abandoning Ship

The final trick relied on by proponents of the plain meaning rule is perhaps the most disquieting. If it looks like the plain meaning of a text is going to take a court where it definitely does not want to go (despite the possibilities offered by artful text selection, adjusting the co-text or switching the meaning of meaning), all is not lost -- the court can always abandon the rule. This is usually done quietly, without drawing attention to the strategic withdrawal, but on occasion it is done with bravura. Here are some examples from judgments written by proponents of the plain meaning rule on the Supreme Court of Canada:

In R. v. McIntosh Lamer C.J. refused to abandon the stark, literal meaning of s. 34(2) of the Criminal Code. He wrote:

... where, by the use of clear and unequivocal language capable of only one meaning, anything is enacted by the legislature, it must be enforced however harsh or absurd or contrary to common sense the result may be (Maxwell on the Interpretation of Statutes, supra, at p. 29). 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.(13)

In R. v.Michaud he relied on purposive analysis to fill a small gap in s. 187 of the Criminal Code which expressly prohibited the opening of sealed evidence packets. In that case he wrote:

... 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.(14)

The following passage is from the judgment of Iacobucci J., speaking for the Court in Canada v. Antosko. In this passage he asserts the primacy of the text:

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....(15)

In Rizzo v. Rizzo, however, Iacobucci J. offers a different analysis. In the following passage, the governing factor is not the text but the intention of the legislature:

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 .... However, with respect this analysis is incomplete.


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 [Act], its object or the intention of the legislature; nor was the context of the words in issue appropriately regarded.(16)

In Canada v. Friesen, and indeed in most cases, Major J. is a staunch supporter of the plain meaning rule. In Friesen he writes:

... 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.(17)

But when the apparent meaning of a text produces results that are truly unacceptable, even he abandons the rule. This is what says in Canada v. Schwartz:

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.(18)

These are not isolated examples. They are the tip of a mighty iceberg. Based on my reading of the relevant case law, I think it is fair to say that the plain meaning rule applies if the text seems plain - unless the judge would rather not apply it, in which case it doesn't apply. This is not my idea of certainty. Nor is it my idea of rule of law.


At first blush, the claim that some texts have a plain meaning seems straightforward, even obvious. But when you look more closely at what proponents of the plain meaning rule do, you discover things like arbitrary text selection, elastic definition of the co-text, shifting and uncertain conceptions of meaning, highly subjective judgements about plausibility, reliance on a discredited view of language, and a lack of commitment to the rule itself. In my view, these practices undermine the credibility of the plain meaning rule and in particular the claim that if the meaning is clear, the outcome is determined by the fixed meaning of the text. There is no such thing as fixed meaning, and even if there were the courts are clearly not bound by it.


Fidelity to Legislative Intent (FLI) Trick # 1 Presume That Nothing Worth Mentioning Has Changed

If legislation is regarded as an ordinary speech act (i.e., a purposeful attempt to communicate a particular meaning), it follows that legislative meaning is fixed once and for all at the time of enactment and that courts, if they want to be faithful to the intentions of the legislature, must seek to discover that original meaning. However, given the role of context in inferring intended meaning, this is not necessarily an easy thing to do. Some statutes were enacted more than a century ago; some were enacted in other jurisdictions; some have been amended many times. The context in which a statute was enacted may bear little resemblance to the context in which it now operates or in which it is to be interpreted. And in truth, the courts lack the resources required to reconstruct and imaginatively inhabit contexts other than their own. The original meaning rule may be rhetorically necessary, but it is often impossible to apply.

This problem is avoided in practice by presuming that the context in which the legislation was enacted is not significantly different from either the context in which it now operates or the context in which it is being interpreted. The burden of proving some significant divergence in these contexts, one that would affect meaning, falls on the person alleging it. In the absence of such an allegation, current meaning is presumed to be the same as original meaning.

This approach works well in practice for it allows the courts to carry out the essential work of adapting ageing statutes to changing circumstance without having to give up the advantages of a fixed legislative intent. The only drawback to the approach is that, like the doctrine of presumed intent examined below, presumed absence of change is not necessarily true. In cases where these presumptions are false, the real work that courts do in interpreting legislative texts is hidden by the fiction.

FLI Trick #2 More Meaning Games

On occasion the original meaning rule is expressly invoked and relied on to justify the outcome in a statutory interpretation case. The courts take the rule seriously, but the concept of meaning relied on for purposes of the rule turns out to be as equivocal in this context as it is in the context of the plain meaning rule. Sometimes "original meaning" refers to the original connotation of the words used by the legislature - their abstract sense or their dictionary definition. At other times "original meaning" refers to their original denotation - the facts to which the words would have been applied when the legislation was first enacted. Obviously the connotation of words is much less dependent on context, and therefore less affected by change, than their denotation. By shifting between these two understandings of "original meaning", the courts are able to ensure appropriate outcomes without appearing to abandon their commitment to legislative intent.

This trick is nicely illustrated by the contrasting judgment of Lamer J. (as he then was) with the judgment of L'Heureux-Dubé J. in Hills v. Canada.(19) The issue in the case was the meaning of "financing" in s. 44(2) of the Unemployment Insurance Act (as it then was). Section 44(1) disqualified claimants from receiving insurance benefits if their loss of employment was due to a labour dispute at their place of work. However, 44(2) created an exception for claimants who could prove that they were not "participating in or financing or directly interested in" the dispute that caused the work stoppage. The court had to decide whether a claimant could be said to be "financing" a dispute if a portion of his regular union dues was paid into a central fund and strike pay was later taken from that fund and paid to workers in a different local who went on strike at his place of work.

L'Heureux-Dubé J. relied on the original denotation of "financing" in concluding that the indirect and involuntary form of payment at issue here was not a form of "financing" within the meaning of the section. As she explained:

The original "financing" provision, enacted in 1935 and re-enacted in 1940, was drafted at a time when very different social conditions prevailed, particularly in the area of labour relations.... In 1935 and until 1944, labour unions were purely voluntary organizations. Individuals would join unions on a voluntary basis and would make their financial contributions in the same manner. They were therefore presumed to be intentionally financing the union's activities within the meaning of the disentitlement provision.(20)

She therefore concluded that the original meaning of "financing" was "giving direct and voluntary financial support" of the sort that would have been contemplated by the legislature when the provision was first enacted in 1935. In other words, she relied on its original denotation.

Lamer J., on the other hand, relied on the original connotation of "financing", which he discovered by consulting the dictionary:

"Financing" means [TRANSLATION] "obtaining the capital necessary to operate" (Petit Robert 1 (1986)), [TRANSLATION] "paying, providing money" (Grand Larousse de la langue française (1973)).(21)

As Lamer J. pointed out, understood in this way the term is broad enough to cover any contribution, voluntary or involuntary, direct or indirect, from one person to another : "it is impossible to find in the word 'financing' used by itself a requirement of active and personal participation or a direct link between the claimant's contribution and the immediate labour dispute". Lamer J. completed his analysis by connecting this broad definition of "financing" to the legislature's original intent:

As L'Heureux Dubé J. has indicated, the wording of s. 44(2)(a) has received little or no alteration since the Unemployment Insurance Act was adopted in 1940.... Though aware of the changes that have occurred in labour relations, Parliament has not felt it necessary to limit the application of a word of general import undoubtedly because it intended to cover all situations to which the word might apply. We cannot assume that this is a mere oversight.(22)

For Lamer J. fidelity to legislative intent means that the court must continue to apply the original connotation of "financing" to whatever facts meet the defining criteria (regardless of consequences) unless and until the legislature signals some other intent.

While both judges were thus faithful to the original meaning of the text to be interpreted, they came up with rather different versions of that meaning. And once again, a rule that purports to constrain the freedom of interpreters turns out on closer analysis to offer a significant choice.

FLI Trick #3 Distinguish Sloppy Drafting From Legislative Error

Fidelity to legislative intent requires that courts correct any errors in a legislative text that are owing to the inattention or lack of skill of the drafter. Such errors must be corrected to bring the text in line with the meaning actually intended by the legislature. However, if the text contains errors or anomalies for which the legislature itself is responsible, fidelity to legislative intent requires that courts do nothing. To cure a legislative oversight would be a form of amendment and an impermissible encroachment on the legislative sphere.

In principle this distinction is coherent enough, and in practice there are some errors for which there is real evidence of a drafter's error. But there is also a large category of errors for which it is impossible to assign responsibility. Errors in this category are up for grabs: depending on its preferred outcome, the court can notionally rewrite the text or piously refuse to touch it -- both in the name of fidelity to legislative intent. These competing possibilities are nicely illustrated in the majority and dissenting judgments of the court in R. v. McIntosh.

The McIntosh case was concerned with the self-defence provisions of the Criminal Code, which read as follows:

[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 ... [no requirement to retreat].

[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 ... [no requirement to retreat].

[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 ...
(c) he declined further conflict ... or retreated from it as far as it was feasible to do ....

The issue was whether an accused who provoked an assault on himself could rely on subs. 34(2). The Crown argued that the less onerous conditions for self-defence set out in subsections 34(1) and (2) were available only to persons who defended themselves from assaults that they did not provoke. In cases where the accused provoked the assault, the more onerous conditions set out in s.35 had to be met. The Crown explained that s.34(2) did not expressly refer to this non-provocation requirement because the reference was inadvertently omitted when the Criminal Code was overhauled in 1955. In short, the absence of the words "not having provoked the assault" in subs. 34(2) was due to a drafter's error.

McLachlin J. (dissenting) accepted this argument. In her view, the legislative evolution of ss. 34 and 35 offered persuasive evidence that the drafter who prepared the 1953-54 version of the Code accidentally changed the meaning of the provisions when all he sought to do was improve the style. She pointed out that the former wording of subs. 34(2) in the 1892 and 1906 versions of the Code subs.34(2) clearly alluded to the non-provocation requirement:

(1) Criminal Code, S.C. 1892

[Self-defence against unprovoked assault]
45. Every one unlawfully assaulted, not having provoked such assault, is justified in repelling force by force, if ... [same as now]; and every one so assaulted is justified, though he causes death or grievous bodily harm, if ... [same as now] he causes it under reasonable apprehension of death or grievous bodily harm from the violence with which the assault was originally made or with which the assailant pursues his purpose, and if he believes, on reasonable grounds, that he cannot otherwise preserve himself from death or grievous bodily harm.

(2) Criminal Code, R.S.C. 1906

[Self-defence against unprovoked assault]
53.(1) Every one unlawfully assaulted, not having provoked such assault, is justified in repelling force by force, if ... [same as now]

[Extent justified]
(2) Every one so assaulted is justified, though he causes death or grievous bodily harm, if ... [same as now]

However, in the 1953-54 Criminal Code, the words "everyone so assaulted" became "everyone who is unlawfully assaulted":

(3) Criminal Code, R.S.C. 1953-54

[Self-defence against unprovoked assault.]
34.(1) Every one who is unlawfully assaulted, not having provoked such assault, is justified in repelling force by force, if ... [same as now]

[Extent justified]
(2) Every one who is unlawfully assaulted and who causes death or grievous bodily harm, is justified if ... [same as now]

It was obvious to McLachlin J. that the drafter incorrectly replaced "so assaulted" with "unlawfully assaulted"; he should have written "unlawfully assaulted, without having provoked the assault".(23)

Lamer C.J., looking at the same evidence, drew a different inference. He thought the 1892 and 1906 versions were both ambiguous, but this ambiguity was deliberately resolved in the 1953-54 version. He wrote:

There is a clear ambiguity in this provision. Does the expression "every one so assaulted" refer to "[e]very one unlawfully assaulted", or to "[e]very one unlawfully assaulted, not having provoked such assault"? This question is academic, since Parliament appears to have resolved the ambiguity in its 1955 revision of the Criminal Code, S.C. 1953-54, c. 51.(24)

Unlike McLachlin J., who blamed the drafter, Lamer C.J. attributed the change in the 1955 revision to Parliament. Since this was an intentional change, it was his duty to give effect to it even though it admittedly was anomalous and led to absurdity:

Even though I agree with the Crown that the interpretation of s. 34(2) which makes it available to initial aggressors may be somewhat illogical in light of s. 35, and may lead to some absurdity, I do not believe that such considerations should lead this Court to narrow a statutory defence. Parliament, after all, has the right to legislate illogically (assuming that this does not raise constitutional concerns). And if Parliament is not satisfied with the judicial application of its illogical enactments, then Parliament may amend them accordingly. (25)

Once again, fidelity to the fixed intention of Parliament proves to be wonderfully malleable. It justifies McLachlin J. in effectively rewriting the provision and it equally justifies Lamer C.J. in refusing to make any change. Like the plain meaning rule, it constrains the rhetoric of the court but not its practice.

FLI Trick # 4 Presume That The Legislature Wants What You Want

Intentionalism's most dazzling trick is the doctrine of presumed intent. To appreciate how effectively this trick has worked, in English Canada at least, all we need do is look at the success of Elmer Driedger's so-called "modern principle" which appeared in the first and second editions of his book Construction of Statutes.(26) Anyone who has read these early editions of Driedger will be struck by their unremitting positivism and commitment to the doctrine of fidelity to legislative intent. In Driedger's garden, judges do not make law; they do not even interpret it; they "construct" it -- a term which suggested to Driedger a workmanlike subordination to legislative plan. His approach is summarized in the modern principle:

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.(27)

Despite my effort to give more meaningful content to this principle in the 3rd edition of Driedger, the Supreme Court of Canada continues to cite and rely on the principle from the 2nd edition. I think it is important to notice what Driedger meant by the intention of Parliament. In the 2nd edition he wrote:

It may be convenient to regard "intention of Parliament" as composed of four elements, namely:

  1. The expressed intention-the intention expressed by the enacted words;
  2. The implied intention- the intention that may legitimately be implied from the enacted words;
  3. The presumed intention- the intention that the courts will in the absence of an indication to the contrary impute to Parliament; and
  4. The declared intention - the intention that Parliament itself has said may be or must be or must not be imputed to it.(28)

By virtue of this definition, the intention of Parliament is deemed to include everything the courts care to impute to Parliament, so long as it does not contradict what Parliament actually said. This is truly a "convenient" way to regard legislative intent.

As defined here, the notion of presumed intention covers not only the standard presumptions (for example, presume that the legislature does not intend to interfere with private property rights) but also evolving presumptions (for example, presume that the legislature does not intend to make women poorer). It also covers the broader and vaguer presumptions (for example, presume that the legislature is reasonable and that it wants to avoid absurdity and injustice). In fact, Driedger's notion of presumed intent has no real limit. It covers whatever values or policies the court sees fit to attribute to the legislature. With this one little trick, Driedger has made the legislature -- our only legitimate law-maker -- intend whatever judges want it to. Unsurprisingly, the courts find it easy to be faithful to this notion of legislative intent.

Why does Driedger, a staunch positivist and a founding father of Canadian intentionalism, do this? Why does he let this snake into his intentionalist garden? The answer is that he has no choice. Actual legislative intent is not enough. In a significant number of cases, the statute does not tell the court what to do and there is no evidence of relevant intention in sources like Hansard. When it appears that the legislature never considered the problem facing the court, when there is no relevant evidence of legislative intent, the court has no choice but to make something up. However, any embarrassment that might flow from such unavoidable bouts of judicial law-making is avoided by presuming that the part made up by judges is in fact what the legislature intended all along.


In my view, the notion of legislative intent is meaningful and has an important role in statutory interpretation. The courts are not free to decide cases on the basis of the state of their digestion or their personal preference. It is sometimes possible to draw compelling inferences about the intended purpose or meaning of a legislative provision and when this happens the court must ordinarily follow the direction of the legislature. The problem with legislative intent is not that it fails to give direction, but that it only goes so far. It is an important consideration but it is not the only consideration. And it often stops short of providing the court with the answers it needs.

The doctrine of fidelity to legislative intent becomes a hindrance rather than a help when the court refuses to acknowledge these limitations. The pretence that judges do not make law, but simply do what they are told tends to undermine judicial credibility. And ironically it undermines the efficacy of legislative intent as a constraint on judicial choice. If anything can be imputed to the legislature, there is no meaningful distinction between legislative and judicial intent. Anything goes.


In my view it is not possible to bridge the gap between judicial practice and judicial rhetoric by making the practice conform to the rhetoric. The plain meaning rule presupposes that particular meanings somehow inhere in texts, but in fact they don't. The doctrine of fidelity to legislative intent presupposes that the legislature has solved every problem, but obviously it did not. There is no point in demanding the impossible from our institutions. Rather, we should respond to unyielding realities by acknowledging them and making the best of them. We should acknowledge that the legislative text and legislative intention are incomplete sources of law and that interpretation is needed to complete the job. We should then focus on what we can reasonably and profitably demand from our interpreters and how these demands can be incorporated into a more serviceable mythology.

Ultimately the gap between rhetoric and reality will not be closed unless we get rid of certain persistent but overly simplistic notions in our mythology. I have in mind in particular the idea that legitimate law is co-extensive with legislation made by elected representatives of the people. Only some of the law that governs us meets this favoured description. Many legally binding rules and regulations are made by bureaucrats and administrators or technicians working for non-governmental organizations. In common law jurisdictions, much private law continues to be made by judges. International conventions, which are negotiated at the international level and concluded under the prerogative, have recently become a major source of domestic control. If there really were no more to democracy than law- making by the elected representatives of the people, we would be in deep political trouble.

In fact democracy is better defined as government in accordance with a constitution of which an elected legislature is one of several necessary institutions. Democracy requires periodic elections, and it equally requires freedom of expression and publication, open government, adequate access to information, a non-independent military, an independent judiciary and the like. To my mind, the resolution of interpretation disputes by judges is a highly democratic exercise in government. The subject herself is allowed to participate directly in the law-making process by arguing about how particular rules should apply to her and her special circumstances. She is entitled to make representations not only about the meaning of the applicable rule, but also about its purpose and consequences and how it should be reconciled with the other rules and norms of the system. She gets to address the decision-maker in the eye and explain her personal interest and perspective. What could be more democratic?

Another unhelpful oversimplification is our tendency to equate the rule of law with the application of pre-fixed rules by a-political automatons. This is not how it works in practice. Those who interpret and apply the law inevitably bring their own context to the job. Short of lobotomy, there is simply no way to disembarrass oneself of one's context. Instead of denying the role of context in interpretation, we should try to understand how it works and what its implications are. If we want to avoid bias and partiality in the application of law, for example, we might consider appointing a more diverse group of persons to our courts and tribunals. Historically judges have been drawn from a very narrow and elite segment of society. Quite naturally, they have perceived their own assumptions to be unbiased truth and protection of their own interests to be impartial protection of the public interest. In this context, challenges by those with different assumptions or interests are readily dismissed as bias or "special" interest. While courts and tribunals have become somewhat more representative in recent years, significant change cannot occur until the legal profession not only lets in members of traditionally excluded groups but also becomes aware of and responsive to the contexts inhabited by those groups.

Finally, we must reject the idea that if courts are not bound by the plain meaning of the text or the intention of Parliament they are free to do whatever they like. Proponents of legal realism and critical legal studies sometimes seem to embrace this position, but it is not tenable in my view. For one thing, it ignores the difference between the legislative text or rule and the law. It is possible for additions to or departures from legislative texts or rules to be justified in a legally acceptable way. In a democracy such a legal justification is necessary and sufficient condition of legitimacy. The outcome of an interpretation dispute is justified in a legally acceptable way if the judge uses techniques that are recognizably legal and appeals to plausible (though not necessarily traditional) legal norms. The judge is free, not to go joyriding, but merely to do the work of interpretation, which consists of trying to understand the text and the intention, working out the implications, adapting rules to varied and changing circumstances and ensuring that important legal values are respected -- both those entrenched in the written constitution and those that are part of the evolving common law.

In short, judges are subject to the constraints imposed by having to justify their decisions to the community they serve in terms of fixed and evolving legal norms. If their community is diverse or in the process of change, they will not satisfy everyone. Some of their decisions will seem controversial to some, or biased, or simply wrong. What I describe here is clearly not rule of law heaven; it does not give us the clarity and certainty that we want. But it is preferable to the illusions of the emperor, which give us nothing at all.


1. [1991] 3 S.C.R. 72.

2. [1996] 2 S.C.R. 171

3. Ibid. at pp. 192-3.

4. McCraw, supra note 1, at p. 80.

5. See R. v. McIntosh, [1995] 1 S.C.R. 686, at pp. 697-98.

6. [1995] 2 1028, at p. 1050. My emphasis.

7. Ibid., at p. 1054. My emphasis.

8. Ibid.

9. Supra note 6, at p. 1055.

10. Ibid., at p. 1081.

11. For an introduction to the massive body of relevant literature, see G. Green, Pragmatics and Natural Language Understanding, 2nd ed. (Mahwah, New Jersey: Lawrence Erlbaum Ass., 1996), at p.10.

12. G. Brown and G. Yule, Discourse Analysis (Cambridge: Cambridge University Press, 1983), at p. 11.

13. Supra note 5, at p. 704.

14. [1996] 3 S.C.R. 3, at pp. 30-1.

15. [1994] 2 S.C.R. 312, at p. 327. My emphasis.

16. [1998] 1 S.C.R. 27, at pp. 40-1.

17. [1995] 3 S.C.R. 103, at pp. 136-37.

18. [1996] 1 S.C.R. 254, at p.

19. [1988] 1 S.C.R. 513.

20. Ibid., at p.

21. Ibid.

22. Ibid., at p.

23. Supra note 5, at p.

24. Ibid.

25. Ibid.

26. E. Driedger, Construction of Statutes, 2d ed. (Toronto: Butterworths, 1983).

27. Ibid., at p. 87.

28. Ibid., at p. 106.

Note: You can reach Ruth Sullivan, the author of this article, via the Webmaster.


Please feel free to contact M. Anne Vespry
if you can suggest other resources.