Article body


On several occasions, the Supreme Court of Canada has suggested that courts should apply what I have elsewhere described as a presumption of restraint when interpreting criminal offences; that courts should presume that Parliament did not intend the offence in question to prohibit conduct that is widely accepted as benign or laudable.[1] This presumption is often expressed as if it were simply an instantiation of the broader principle that legislation should not be interpreted in an “absurd” fashion. Construed in such a way, however, the presumption is in grave tension—or flatly inconsistent—with other foundational separation of powers principles and canons of statutory interpretation. Chiefly, it might seem to suggest that it is open to the courts to narrow the scope of criminal offences simply on the basis that they regard them as overinclusive on policy grounds. Such an approach would be difficult to square with the notion of parliamentary sovereignty, or the idea that statutes should be interpreted in a purposive manner. Yet, in the modern era, the Supreme Court has consistently taken the view that criminal offences should be construed with ideas like parliamentary sovereignty front and center. This focus is difficult to reconcile with how the presumption of restraint is often expressed. Indeed, these ideas have been taken so seriously by the Supreme Court that competing canons—like the “rule” of strict construction—have largely fallen by the wayside.

In this paper, I set out to provide an account of the presumption of restraint that could sit comfortably in our constitutional system of arrangements. This account draws upon an underappreciated aspect of the work of Lon Fuller: in particular, his analysis of the role that customary law plays in modern legal systems. Given the legislature’s intention to use general rules to guide citizens, Fuller suggests, it must be taken to have anticipated that—in the absence of unequivocal language to the contrary—they will construe those rules in a manner that is consistent with widely accepted practices and modes of interaction. As the parenthetical proviso suggests, the presumption can be rebutted with interpretive evidence suggesting a legislative intention to displace the practice. The fact that it is rebuttable, and exists in order to further the legislature’s own intention to guide citizens, means that the presumption is compatible with both parliamentary sovereignty and the purposive approach to statutory interpretation. Relatedly, it accommodates widely shared intuitions that the substantive criminal law, at least sometimes, aims at transforming social norms, and not just reflecting them. Moreover, the considerations that have driven the canon of strict construction into disuse do not apply to the presumption of restraint.

I. The Presumption of Restraint

Appellate courts have frequently held that the ambit of criminal offences is more restricted than a plain reading of their text would suggest. They have done so, often, on the basis that it would be “absurd” to suppose that Parliament intended to criminalize certain widely accepted practices or modes of interaction. In Munroe, for example, the Ontario Court of Appeal gave a narrow reading of then paragraph 171(1)(c) of the Criminal Code.[2] That provision stated: “Every one who ... (c) loiters in a public place and in any way obstructs persons who are there ... [i]s guilty of an offence”. The Crown urged a broad reading of “loiters”, as well as the phrase “in any way obstructs”. Justice Cory (then at the Ontario Court of Appeal) rejected that suggestion, arguing that the Crown’s interpretation would mean that various ordinary courses of action, widely regarded as benign or harmless, were now criminal:

If the Crown is correct in its contention a good many members of the community must have breached its provisions. For example, a spouse waiting for his or her tardy mate must, of necessity, loiter about the appointed meeting place and, in so doing, obstruct in some degree the passage of others. Indeed, one can well imagine that the appointed meeting place was the exit to the tunnel from Union Station to the Royal York Hotel. A missed GO train by one of the spouses may result in the other spouse waiting at least 20 minutes and perhaps an hour with nothing better to do than wander aimlessly about during that time span.

He continued:

The window-shopper, with time to kill, the devout member of a religious group earnestly seeking to proselytize by handing out pamphlets, the Salvation Army member charitably seeking funds for the welfare of the needy, the anxious and ambitious politician seeking to shake hands with passers-by and to press home his point of view — all these would be caught under the Crown’s interpretation of the section.[3]

The Supreme Court’s decision in Skoke-Graham is also instructive.[4] There, the Court was confronted with then subsection 172(3) of the Criminal Code, which prohibited the willful disturbance of the order or solemnity of an assemblage of persons meeting for religious worship. Justice Dickson (as he then was) noted: “Parliament could not have intended that [the subsection] could be triggered by conduct which is not disorderly or productive of disorder.” In ordinary usage, “disturbance” has a potentially wide meaning. Nonetheless, Justice Dickson observed that, if the term “disturbance” were given such a broad reading, “a man might be convicted ... for failing to take off his hat in a church, or failing to keep it on in a synagogue.”[5] He chiefly grounded his interpretation in textual considerations—in particular, the heading for then sections 169 to 175 of the Criminal Code.[6] But his other remarks suggest that the presumption of restraint was also at work.

Later, in Hinchey, the Supreme Court interpreted paragraph 121(1)(c) of the Criminal Code.[7] That provision states:

Every one commits an offence who ... being an official or employee of the government, directly or indirectly demands, accepts or offers or agrees to accept from a person who has dealings with the government a commission, reward, advantage or benefit of any kind for themselves or another person, unless they have the consent in writing of the head of the branch of government that employs them or of which they are an official...[8]

The entire Court agreed that the emphasized portion of the provision seems to capture a quite broad range of conduct, including conduct which would strike many of us as relatively innocuous. Justice Cory, writing for three judges, observed: “[I]f a government employee accepts, on a rainy day, a ride downtown from a friend who does business with the government he has received a benefit. That could hold true as well for the cup of coffee or occasional lunch bought by the friend for the government employee.”[9] Yet the whole panel agreed that it would be unacceptable to interpret paragraph 121(1)(c) in such a way that it criminalized the acceptance of a cup of coffee. Partly on that basis, a narrower reading was preferred.

Finally, the Supreme Court has, on a number of occasions, construed the assault and sexual assault provisions of the Criminal Code in a relatively narrow fashion. In Cuerrier,[10] a majority refused to read “fraud” in paragraph 265(3)(c) of the Criminal Code—a provision setting out the circumstances under which consent to applications of force is vitiated—in such a way that it required proof of mere deception. To adopt that reading, Justice Cory held, would effectively criminalize many acts of deception that “lack the reprehensible character of criminal acts.”[11] Lies about one’s age, for example, would be criminal so long as they induced the complainant to give apparent consent to sexual acts. Justice Cory continued: “The same result would necessarily follow if the [defendant] lied as to the position of responsibility held by him in a company; or the level of his salary; or the degree of his wealth; or that he would never look at or consider another sexual partner; or as to the extent of his affection for the other party; or as to his sexual prowess.” He concluded with a rhetorical question: “The lies were immoral and reprehensible but should they result in a conviction for a serious criminal offence? I trust not.”[12]

Justice McLachlin (as she then was) similarly observed that, if mere deception constituted “fraud” within the meaning of subsection 265(3), many quotidian acts of apparently consensual touching, such as handshakes and back-slaps, would be assaults:

[T]his approach vastly extends the offence of assault. Henceforward, any deception or dishonesty intended to induce consent to touching, sexual or non-sexual, vitiates the consent and makes the touching a crime. Social touching hitherto rendered non-criminal by the implied consent inherent in the social occasion — the handshake or social buss — are transformed by fiat of judicial pen into crimes, provided it can be shown that the accused acted dishonestly in a manner designed to induce consent, and that the contact was, viewed objectively, induced by deception. No risk need be established, nor is there any qualifier on the nature of the deception. Will alluring make-up or a false moustache suffice to render the casual social act criminal? Will the false promise of a fur coat used to induce sexual intercourse render the resultant act a crime? The examples are not frivolous, given the absence of any qualifiers on deception.[13]

In the view of everyone on the Court, other than Justice L’Heureux-Dubé, Parliament could not have intended to give the offence of assault such a wide ambit.[14]

Later, in JA, Justice Fish articulated concerns similar to those animating the majority and concurring reasons in Cuerrier.[15] Justice Fish took issue with the JA majority’s conclusion that subsection 273.1(1), which sets out the circumstances under which there is consent to sexual touching, precludes advance consent. His objection was based, in large part, on the idea that the majority’s interpretation would lead to “absurd” results:

The approach advocated by the Chief Justice would ... result in the criminalization of a broad range of conduct that Parliament cannot have intended to capture in its definition of the offence of sexual assault. Notably, it would criminalize kissing or caressing a sleeping partner, however gently and affectionately. The absence of contemporaneous consent, and therefore the actus reus, would be conclusively established by accepted evidence that the complainant was asleep at the time. Prior consent, or even an explicit request — “kiss me before you leave for work” — would not spare the accused from conviction.[16]

The majority rejected Justice Fish’s analysis, for reasons we will see shortly. For now, it is enough to note that Justice Fish was drawing on an accepted line of reasoning. Indeed, we can see this argumentative strategy at work in unlikely places. The Supreme Court’s decision in Jobidon, for example, stands for the proposition that apparent consent to applications of force may be vitiated on public policy grounds.[17] Yet, in the course of explaining how it could be open to the courts to effectively expand a criminal offence, Justice Gonthier observed that the reference to “consent” in section 265 could not be taken at face value:

Assault has been given a very encompassing definition ... If taken at face value, this formulation would mean that the most trivial intended touching would constitute assault. As just one of many possible examples, a father would assault his daughter if he attempted to place a scarf around her neck to protect her from the cold but she did not consent to that touching, thinking the scarf ugly or undesirable ... That absurd consequence could not have been intended by Parliament.[18]

Moreover, Justice Gonthier’s opinion consistently proceeded on the basis that consent to certain rough sporting activities—for example, hockey—would not be vitiated on public policy grounds.[19]

Though a Charter case, rather than a straightforward statutory interpretation decision, the Supreme Court’s ruling in Canadian Foundation for Children, Youth, and the Law is also instructive. There, the majority rejected suggestions that the corrective force defence should be struck down as incompatible with section 7 of the Charter. Chief Justice McLachlin, writing for the majority, claimed that such an approach “would criminalize force falling far short of what we think of as corporal punishment, like placing an unwilling child in a chair for a five-minute ‘time-out’.”[20]

Other examples abound. In Khawaja,[21] the Court construed section 83.18 of the Criminal Code, which made it an offence to “knowingly participate in or contribute to, directly or indirectly, any activity of a terrorist group, for the purpose of enhancing the ability of any terrorist group to facilitate or carry out a terrorist activity”.[22] On a plain reading, one could commit the offence by engaging in “essentially harmless” activities: “For example, a person who marches in a non-violent rally held by the charitable arm of a terrorist group, with the specific intention of lending credibility to the group and thereby enhancing the group’s ability to carry out terrorist activities.”[23] The Court found that “the context makes clear that Parliament did not intend for the provision to capture conduct that creates no risk or a negligible risk of harm.”[24] In Déry, the Supreme Court took the view that subsection 24(1) of the Criminal Code does not apply to the offence of conspiracy—finding that there is no offence of “attempted conspiracy”—on the basis that Parliament did not intend to create “thought crimes.”[25] And though, in practice, little tends to turn on whether a causal contribution is “outside the de minimis range,” the Smithers test also reflects a concern that causation requirements should not be read in an over-expansive way.[26]

Whole criminal law doctrines arguably reflect a concern for the effect that a “plain reading” of criminal offences would have on practices and courses of action that are widely regarded as benign, salutary, or insufficiently blameworthy. For example, criminal offences are presumptively read in such a way that liability attaches to acts rather than (pure) omissions. Criminal Code provisions imposing positive legal duties upon discrete groups have, by and large,[27] been construed narrowly.[28] Thus, in Browne, Justice Abella (then at the Ontario Court of Appeal) interpreted the undertakings provision of the Code in a narrow fashion, lest it be construed in such a way that mere promise breaking could give rise to criminal liability.[29] There is a compelling argument that the basis for the courts’ hesitation to construe positive legal duties in an expansive manner lies in the intuition that, all other things being equal, the mere failure to intervene on behalf of a third party is not per se regarded as blameworthy.[30]

Before moving on, it is worth taking a moment to distinguish the presumption of restraint from doctrines and cases that bear a superficial similarity. In particular, it might be thought that the “doctrine” of de minimis non curat lex provides a salient example of the presumption of restraint. According to that doctrine, trial judges may refuse to apply criminal offences where the conduct in issue amounts to a merely technical or “trifling” violation.[31] Thus, it has been applied in cases where the defendant stole goods of vanishingly little value or possessed trace amounts of narcotics.[32] The de minimis doctrine, however, even if it exists in Canada,[33] is not a canon of statutory interpretation. It is invoked as a defence in cases where it is clear that the defendant’s conduct falls within the ambit of the offence in question, but where the violation strikes the trial judge as no more than technical in nature.[34] The presumption of restraint, by contrast, is used to interpret the offence by way of determining what its ambit is in the first place.

It might seem that the Supreme Court of Canada’s decision in Labaye provides an illustration of the presumption at work.[35] There, the majority purported to construe the offence of criminal indecency on the basis that it encompasses only “harmful” acts, thereby reducing its ambit. But I would draw several distinctions between the reasoning in Labaye and that used in the presumption-of-restraint line of authorities. First, the latter tends to invoke specific courses of action, which Parliament ostensibly “could not” have intended to criminalize, by way of explaining why the offence in question must be understood more narrowly than its plain meaning would otherwise suggest. In Labaye, by contrast, the majority purported to interpret the offence of criminal indecency in light of highly abstract notions of “harm” and “the proper functioning of society” that—they freely acknowledged—would need to be concretized and given content over a long series of cases “[i]n the tradition of the common law.”[36] Second, and relatedly, it is not at all clear that the reasoning in Labaye rests upon ordinary intuitions about the non-criminal nature of this or that course of action, so much as it rests upon constitutional norms and values.[37] The emphasis in Labaye is not on common sense, but on the “fundamental value[s] reflected in” positive laws, such as “our society’s Constitution or similar fundamental laws, like bills of rights.”[38]

The reasoning in Labaye arguably turns the presumption of restraint on its head. In Labaye, the majority begins with the premise that the offence of criminal indecency can capture only harmful conduct that is incompatible with the proper functioning of society, and then goes on to identify (in quite broad terms) circumstances under which that test may be satisfied. It proceeds, in other words, on the tacit basis that the offence captures nothing unless and until one can show that it captures something. The presumption-of-restraint line of authorities moves in the opposite direction: it proceeds on the basis that the offence in question would have an expansive ambit, consistent with its plain language, but for the fact that Parliament could not have intended to criminalize this or that course of action.

For these reasons, I am uncomfortable with the suggestion that Labaye is an illustration of the presumption of restraint. The two are loosely related. But the differences are significant enough that we should set Labaye aside in the discussion that follows.

II. Strict Construction by Another Name?

In a range of cases, then, appellate courts have interpreted criminal offence provisions in a manner more narrow than their plain language would seem to permit, on the basis that the plain reading would catch courses of action that Parliament could not have intended to target.[39] How should we understand this line of authorities?

We should resist the temptation to read these cases as instances in which the courts have applied the canon of strict construction. The rule was described by Justice Dickson (as he then was) in Marcotte:

It is unnecessary to emphasize the importance of clarity and certainty when freedom is at stake. No authority is needed for the proposition that if real ambiguities are found, or doubts of substance arise, in the construction and application of a statute affecting the liberty of a subject, then that statute should be applied in such a manner as to favour the person against whom it is sought to be enforced. If one is to be incarcerated, one should at least know that some Act of Parliament requires it in express terms, and not, at most, by implication.[40]

The canon is animated by two rationales. The first focuses on fairness to the criminal defendant. Insofar as one faces a deprivation of liberty, some suggest, it is important that the deprivation be traceable to a clear decision by Parliament to impose it, and not to the ad hoc decision of a judge. As Justice Wilson put it, the “seriousness of imposing criminal penalties of any sort demands that reasonable doubts [as to the scope of an offence] be resolved in favour of the accused.”[41] The second focuses on fairness to the citizen. To the extent that one wants to make plans, it is important to know in advance which courses of action are criminal and which are not. The canon of strict construction ostensibly guarantees fair notice by ensuring that only conduct falling squarely within the offence in issue may give rise to criminal sanctions.[42]

In some respects, the canon is linked to what I have called the presumption of restraint. The Supreme Court has sometimes linked the presumption of restraint to the need for certainty and fair notice.[43] Moreover, the Court has sometimes tied the presumption to the need to ensure that the criminal sanction is applied only to truly “reprehensible conduct”—a rationale which might appear grounded in fairness considerations broadly resembling those underpinning the canon of strict construction.[44] But the two interpretive techniques are distinct. Note, first, that the canon of strict construction applies only once the interpreting court has already concluded that the provision in issue is irreducibly ambiguous—that is, that the ambit of the offence cannot be settled using ordinary textual and purposive analysis.[45] Where the interpretive question can be resolved without looking to the canon of strict construction, the court will do so.[46] In Hasselwander, Justice Cory stated:

[T]he rule of strict construction becomes applicable only when attempts at the neutral interpretation suggested by s. 12 of the Interpretation Act still leave reasonable doubt as to the meaning or scope of the text of the statute. ... [T]his means that even with penal statutes, the real intention of the legislature must be sought, and the meaning compatible with its goals applied.[47]

Likewise, in Bell ExpressVu, the Court remarked that the canon “only receive[s] application where there is ambiguity as to the meaning of a provision.”[48] Moreover, the Court has made it clear that, for an “ambiguity” to exist, there must be “two or more plausible readings, each equally in accordance with the intentions of the statute”—emphasizing that a finding of ambiguity presupposes that the textual and purposive analysis has already taken place.[49] In Bell ExpressVu, the Court observed that a provision is not “ambiguous” for the purposes of the canon merely because reasonable people disagree about the significance of the text and the precise nature and interpretive implications of the legislative purpose.[50] Over and over again, the Court has treated the canon as a rule of “last resort.”[51] The practical effect of this approach has been to make the canon of strict construction all but irrelevant as courts attempt to make sense of criminal offence provisions.[52]

By contrast, the presumption of restraint is brought to bear as one aspect of purposive analysis. It is not invoked as a tiebreaker only once the interpreting court has already determined that the offence provision is ambiguous. Rather, it is invoked by way of discerning what conduct Parliament “could not” or “must” have intended to target when it enacted the offence. The presumption thus operates much (though not exactly)[53] like the presumption of subjective fault, at the front end of the interpretive process.[54] When deployed, the presumption of restraint relies on the notion that it would be absurd (not) to attribute certain intentions to Parliament, and not that a given construction should be imputed to the offence in question notwithstanding legislative intent. Put another way, the presumption is invoked by way of explaining why, despite the plain meaning of the statutory text, Parliament cannot be taken to have intended to target as expansive a range of conduct as it would appear at first glance.

Second, and relatedly, the canon of strict construction is driven by fairness concerns vis-à-vis the defendant. The rationale for the canon is that, faced with a genuinely ambiguous provision, a citizen is entitled to proceed on the basis of the narrower reading—or at least that it would be unfair to punish her for having so proceeded. By contrast, the presumption of restraint is not animated by fairness considerations, but by the need to construe offence provisions in a manner consistent with Parliament’s intentions. A court may apply the presumption, giving the offence in question a narrow interpretation, yet conclude that the defendant nonetheless falls within the ambit of the circumscribed offence. Thus, the presumption has been successfully invoked in a number of cases in which the Supreme Court either overturned the defendant’s acquittal or upheld his or her conviction.[55] Furthermore, the presumption is, as we will see, rebuttable. It is a defeasible heuristic that ostensibly sheds light on Parliament’s intentions, meaning that it may be invoked without necessarily leading to the conclusion that the offence in issue should be construed narrowly at all—that is, so long as there is other interpretive evidence suggesting that Parliament intended the offence to have a broader scope.

As I hinted earlier, the canon of strict construction has largely fallen out of use. In part, this is because the underlying rationale for the rule has eroded. The canon was devised at a time when there were many capital crimes in English law. There is now less justification for giving offences a construction at odds with legislative intent.[56] This development also reflects the now universal acceptance of the principle—itself articulated by Parliament in section 11 of the Interpretation Act—that “[e]very enactment shall be deemed remedial, and shall be given such fair, large and liberal construction and interpretation as best ensures the attainment of its objects”.[57] The courts have recognized that the canon of strict construction is consistent with purposive interpretation so long as it is applied only as a “last resort.” The practical effect of this move, however, has been to make the rule moribund (or all but).[58] By contrast, the presumption of restraint has thrived, precisely because ostensibly it is not only consistent with the purposive approach, but part and parcel of it.

III. The Paradoxical Presumption

So far, I have emphasized remarks suggesting a presumption of restraint on the part of the Supreme Court and other appellate courts. But why only a presumption? Why not say, instead, that the courts have, in the cases I have discussed, imposed a hard substantive constraint on Parliament’s criminal law-making authority? One could, for example, draw upon some version of the harm principle (or other “wrongness constraint”) by way of claiming that there are moral[59] and perhaps constitutional limits on that authority.[60] Indeed, aspects of the Supreme Court’s decisions in Butler and Labaye, as well as the Reference re Assisted Human Reproduction Act,[61] could be read as supporting the conclusion that such limits exist.[62]

My answer is threefold. First and foremost, such a conclusion would be utterly unsupported by the language used by the courts themselves. In the cases discussed in Part I, there is no suggestion whatsoever that Parliament could not criminalize the activities referenced therein—that is, that Parliament could not criminalize mere lies, or the purchase of cups of coffee, etc. Indeed, in most of those cases, such remarks would be pure obiter, since the constitutionality of the offences in question was not in issue, and there was no claim that the presumption of constitutionality should be applied by way of resolving an otherwise intractable interpretive dilemma.[63] These were cases in which the courts were ostensibly engaging in ordinary purposive statutory interpretation.

Second, the Court has sometimes been explicit that it is open to Parliament to criminalize courses of action heretofore thought unobjectionable or insufficiently wrongful. This point was made, for example, by the majority in Hinchey:

The notion of criminality ... is not a static one, but one which very much changes over time. As society changes, the conception of what types of conduct can properly be considered criminal also evolves. There are a myriad of different activities which at one point in time were considered legal, but which we now consider criminal. The offence of criminal harassment is one obvious example. For many years, it was not recognized as criminal to persistently follow someone and cause them to fear for their safety, so long as no contact was made. Now, that has distinctly changed with the addition of s. 264 of the Code, which makes this conduct a crime.[64]

In Beatty, too, a majority held that the actus reus of dangerous driving does not require proof that the defendant’s driving was a “marked departure” from standards of driving expected of reasonable people.[65] This surely suggests that certain driving practices may be regarded as criminal, though they are common, perhaps even ubiquitous. Kent Roach, discussing Beatty, observed: “[T]his approach ... recognize[s] the traditionally dominant role accorded to the legislature in defining the criminal act. There is a danger that reading in a requirement of a marked departure into every criminal act will alter the clear intent of the legislature in defining the criminal act.”[66]

The offence of sexual assault is also illustrative. In JA, the defendant argued that the offence of sexual assault should not be construed in such a way that it would capture the ostensibly “innocent” act of kissing one’s sleeping spouse.[67] The majority rejected that argument, largely on the basis that, even if an expansive understanding of the offence led to “unrealistic” results, that simply reflected legislative intent. Chief Justice McLachlin stated:

In the end, we are left with this. Parliament has defined sexual assault as sexual touching without consent. It has dealt with consent in a way that makes it clear that ongoing, conscious and present consent to “the sexual activity in question” is required. This concept of consent produces just results in the vast majority of cases. It has proved of great value in combating the stereotypes that historically have surrounded consent to sexual relations and undermined the law’s ability to address the crime of sexual assault. In some situations, the concept of consent Parliament has adopted may seem unrealistic. However, it is inappropriate for this Court to carve out exceptions when they undermine Parliament’s choice. In the absence of a constitutional challenge, the appropriate body to alter the law on consent in relation to sexual assault is Parliament, should it deem this necessary.[68]

In effect, the majority held that, barring a constitutional challenge, it is neither here nor there that the offence in question catches courses of action widely regarded as benign or blameless if, ultimately, one concludes that Parliament intended to target that sort of conduct.

My third point is that, from a separation of powers perspective, it would be strange if the courts could narrow the ambit of a criminal offence, irrespective of Parliament’s intentions, solely on the basis that it would otherwise catch courses of action that judges regarded as harmless or insufficiently blameworthy. Even in the Charter context, the Supreme Court has never struck down, or read down, a criminal offence simply on the basis that it targets a course of action—that is, has an actus reus—that is insufficiently wrongful.[69] Where the Court has found actus reus elements unconstitutional, it is because they either infringed some discrete Charter right (e.g., freedom of expression) or were contrary to section 7 principles of instrumental rationality.[70] The majority in Malmo-Levine expressly denied that criminal offences must target discernible other-regarding harms in order to pass constitutional muster. Insofar as the Court has suggested otherwise, it has adopted an extremely elastic vision of what constitutes harm.[71] With all this in mind, it would be odd if judges regarded themselves as free, in non-constitutional cases, to rewrite criminal offences to fit their (possibly idiosyncratic) notions of what courses of action it would be absurd to criminalize.

The better reading of the cases discussed in Part I, then, is that the presumption of restraint is just that: a presumption. It is a heuristic device designed to shed light on what Parliament intended to criminalize, not a quasi- (or pseudo-) constitutional rule that constrains what Parliament can criminalize. It is not, by itself, dispositive of the interpretive issue before the court. On the contrary, the presumption is a conversation starter rather than a stopper: by directing the decisionmaker to presume, in the absence of interpretive evidence to the contrary, that Parliament did not intend to target this sort of conduct, the presumption effectively counsels her to ask whether such evidence exists, and how weighty it is. In this way, the presumption (again, like the presumption of subjective fault)[72] helps to structure the inquiry into legislative intent, without foreclosing any particular outcomes.

What sort of interpretive evidence might rebut the presumption? Most obviously, we might look for express statutory provisions “declaring” that the conduct in question falls within the ambit of the offence in question.[73] The wider context might also strongly suggest that the offence must encompass courses of action which we might otherwise regard as innocuous. For example, there may be other offences that clearly overlap with the offence in issue; since Parliament is presumed not to have intended to create a wholly redundant crime, the only reasonable conclusion may be that the legislature intended to target a different or broader range of activities.[74] Alternatively, Parliament may have explicitly recognized discrete exemptions or immunities that only make sense if one supposes that the offence otherwise encompasses activities traditionally regarded as benign or insufficiently blameworthy. Looking beyond statutory text and context, it may be clear from preambles and the legislative history (including statements made by Members of Parliament) that there was an intention to reform societal attitudes to hitherto accepted activities—or at least acceptance that those activities would be encompassed by the offence in question, if only as a foreseeable side effect of addressing the contemplated “mischief.”[75] In short, we would look to the very sorts of interpretive evidence used in other statutory interpretation cases: text, context, and extrinsic evidence of legislative purpose.

IV. The Common Law Constitution?

To emphasize the modesty of the presumption of restraint is not (without more) to deny that it may reflect and, with important qualifications, protect fundamental values.[76] It is only to stress that the presumption imposes “soft” rather than “hard” barriers. Kent Roach has persuasively argued that interpretive canons and presumptions provide a means by which the courts can articulate foundational values—throwing “speed bumps” in the path of legislatures which might otherwise run roughshod over them—while paying due fealty to the sovereignty of Parliament, which can always respond with a “clear statement” that it intends to override them.[77] In this sense, Roach claimed, clear statement rules, such as the canon of strict construction, provided opportunities for “dialogue” between courts and legislatures long before the emergence of the Charter.[78] Insofar as the courts can give effect to quasi-constitutional values without declaring any hard limits on legislative authority, ultimately leaving the difficult political choices to the elected branches, they may avoid (or at least mute) charges of judicial activism. Writing in the American context, Sunstein and Bickel likewise emphasized clear statement rules as a means by which judges may articulate values without shoving aside the elected branches.[79]

So it is possible to see common law presumptions as reflecting fundamental values “imposed” by the courts on legislative branches.[80] And there is a temptation to understand the presumption of restraint in this way—in particular, to view it as a means by which the courts may forestall the aggressive expansion of the substantive criminal law in an age of rampant over criminalization.[81] As we have seen, the Supreme Court has often referenced the need to reserve the criminal law for truly “reprehensible” conduct and linked the presumption of restraint to this goal.[82]

But I have misgivings about this reading of the authorities. In part, my concerns have to do with the Charter case law I have mentioned. The Supreme Court has, by and large, refused to recognize hard constitutional limits on what Parliament can criminalize, intervening only when a free-standing Charter right has been infringed or when there has been an absence of means–ends rationality.[83] Otherwise, the Court has left Parliament a free hand to determine what courses of action warrant criminal condemnation and punishment.

Now, the fact that the Supreme Court has taken this approach in the Charter context does not logically preclude it from imposing “restraint” on Parliament in the interpretation context. In the latter context, there is neither a threat of invalidation nor anything stopping Parliament from re-enacting the legislation (this time with a clear statement about its ambit). A court might well think itself freer to assert restraint-based values, in the course of interpretation, that it would hesitate to enforce by way of resolving a Charter challenge.

We should not, however, understate the kind of damage that aggressive “interpretation” of a statute can do to a government policy or legislative scheme, or minimize the separation of powers implications of that step. Requiring Parliament either to acquiesce to a statutory regime that fails to achieve its intended objectives, or else re-enact the statute, represents a significant intrusion on the legislative role.[84] Indeed, a declaration of invalidity is modest by comparison, since it relieves Parliament of the burden of living with a statutory provision or scheme it may not have intended to make. For this very reason, the Court in Ferguson observed that constitutional infirmities with a statutory scheme should be remedied with a declaration rather than by reading in or reading down.[85]

This sort of interference may be palatable when the interpreting court is faced with a genuinely ambiguous statutory provision. Where a person’s liberty is at stake and there is no one interpretation that should clearly prevail over another, it is not obviously disrespectful to Parliament to give the benefit of the narrower reading to the defendant. Hence, it is no accident that the canon of strict construction is triggered only after the court has concluded that the offence in issue is ambiguous, having regard to the text and the legislative purpose. As we have seen, however, this reasoning has no application to the presumption of restraint as it has been deployed by the courts. The presumption is mobilized at the outset of the interpretation process, driving the inquiry, ostensibly because it sheds light on what Parliament intended. If this is a judicial imposition of quasi-constitutional values, it is a particularly aggressive example.

That brings me to just what is being presumed here: “restraint.” The presumption has been invoked by way of construing offence provisions as narrower than their plain meaning might otherwise suggest. This can be a deeply controversial—deeply political—approach to take.

When Parliament creates a criminal offence, it purports to set the basic terms of social interaction in the community; to draw lines between the permissible, the mandatory, and the forbidden. Sometimes, perhaps often, those lines reflect norms and values that are already dominant in the community as a whole; Parliament’s aim, in those instances, is simply to reinforce those norms. But it may wish to assume a leadership role and use its criminal law power to change existing social norms, practices, and values—or to settle widespread disagreements within the community as to the permissibility of a course of action. The presumption of restraint may undermine this sort of reformist agenda.

Consider the sexual fraud line of authorities I discussed in Part I. The majority in Cuerrier construed “fraud” in subsection 265(3) in a narrow fashion, drawing on the presumption of restraint.[86] Arguably, neither the presumption nor the resulting interpretation reflected Parliament’s intention to transform social norms regarding how it is appropriate to obtain sexual consent. Certainly, L’Heureux-Dubé’s minority opinion in Cuerrier was indignant, not least because she regarded the majority’s “interpretation” as out of step with the reformist objectives of Parliament.[87] Likewise, the majority opinion in Hutchinson, which implicitly proceeded on the same basis as Cuerrier—that Parliament did not intend to criminalize widely accepted tactics in sexual seduction—was vigorously criticized by the concurring judges on the basis that subsection 273.1(1) was intended to give new recognition to women’s sexual autonomy and, to that end, was intended to disrupt existing social norms.[88] Notably, the Court’s decision in Ewanchuk made no mention of the presumption of restraint,[89] and the majority decision in JA emphasized that the presumption had been rebutted.[90]

Beyond the sexual assault context, we can see other instances in which the presumption of restraint might appear to push against legislative attempts at social reform. Take, for example, the recent decision in Jarvis, in which the Court interpreted the offence of voyeurism.[91] The Court did not expressly reference the presumption of restraint, but the majority was clearly concerned that a narrow understanding of “reasonable expectation of privacy”—one reflecting only traditional conceptions of “criminal” conduct—would fail to give effect to Parliament’s intention to protect sexual autonomy in an era of rapid technological change.

The mere fact that the presumption of restraint fails to reflect Parliament’s intentions in some cases is not especially striking, or even interesting. Rules are, by definition, overinclusive,[92] and the presumption is rebuttable. For now, I only want to draw attention to the deeply political nature of the presumption, which is at odds with a widely held instinct about the modern role of Parliament and the criminal law—namely, that it should not just reflect existing social norms, but actively seek to shape them.[93]

V. The Internal Morality of Criminal Offences

With all this in mind, we should hesitate to see the presumption of restraint as an exercise in which judges aggressively foist values upon an unwilling Parliament. Given the highly political nature of the presumption, and the fact that it is in grave tension with other strains of criminal and constitutional law over the last thirty years, I want to propose an alternative possibility. (I put it no more strongly than that.) Specifically, I suggest that the presumption finds its footing in the very idea of what it means to craft a criminal offence in the first place.

A. Fuller and the Internal Morality of Self-Executing Guidance

In The Morality of Law, Lon Fuller claimed that any putative lawmaker must, logically, adhere to eight principles of legal “craftsmanship.”[94] Failure to abide by these principles does not merely mean that one has made “bad law”; it means that one has failed to make law, properly speaking, at all. After all, the law exists first and foremost to provide guidance for citizens, who by and large will be expected to follow and apply its directives themselves. When a putative law-maker fails to publicize the laws it purports to create, or makes them retroactive or incomprehensible, she has failed to create something that citizens are able to follow or apply, and which therefore fails as a guide. This is true, moreover, whether or not the law-maker’s objectives are just or unjust—that is whether or not the laws are intended to guide subjects in good or bad ways, or towards good or bad ends. Whatever the law-maker’s intentions, if she resorts to law—in the sense of general directives—she must mean to guide citizens in some way and, to the extent she has crafted putative laws that are incapable of discharging that function, has failed as a law-maker. In this sense, Fuller claimed that there is an “internal morality” of law: the very idea of law-making presupposes rules of craftsmanship that are responsive to the agency of subjects; that “speak” to them as active law-users rather than passive law-takers, and that respond to their needs as planning agents.

How far Fuller’s account actually goes in drawing a necessary connection between law and morality is a matter of debate.[95] For my purposes here, it does not matter. It is enough to observe that, inasmuch as one proceeds on the basis that criminal offences are instruments intended to provide guidance to citizens, that modest starting point has implications for how courts ought to read them. It leads, for example, to the interpretive presumption that Parliament did not intend criminal offences to have retroactive effect, as well as the presumption that it intended the text of the offence to be given its ordinary meaning.[96] More controversially, it also leads to the conclusive presumption that criminal offences do not purport to guide people under circumstances in which they are unable to control their actions, and to the rebuttable presumption that true crimes require proof of subjective fault.[97] These interpretive presumptions follow from the basic idea that, when Parliament creates a criminal offence, it purports to craft directives that will be applied by the citizens themselves.

The presumption of restraint may also find its grounding in logical propositions about what it means for a criminal offence to offer guidance, and not in moral propositions imposed by judges upon recalcitrant legislatures. Understood in this way, the presumption does not reflect any moral or political judgment about the sorts of conduct Parliament ought to criminalize; it reflects only the respectful judgment that, whatever Parliament intended to criminalize, it perforce intended to guide citizens in some fashion. It does not, to that extent, entail any separation of powers difficulties.

B. The Theoretical Shallowness of the Presumption

I begin with the observation that when courts rely upon the presumption of restraint, they do not tend to explain just what it is about this or that course of action that presumptively makes it an absurd target of criminalization. When the Supreme Court finds it absurd to suppose that Parliament intended to criminalize “seductive” lying, or buying cups of coffee for a friend, or putting a scarf around the neck of one’s own child, it does not provide—or purport to provide—any kind of a priori moral argument that these courses of action are innocent or only trivially blameworthy. Nor does it draw upon any kind of political argument by way of explaining why these courses of action should be regarded as presumptively beyond the limits of the criminal law. Rather, the Court makes a bare appeal to the moral intuitions of its readers that these activities “must” fall outside those boundaries, at least presumptively.

The temptation for many will be to see this preference for shallow—in the sense of theoretically unambitious—reasoning through a Sunsteinian lens.[98] It would be unwise for, say, the Supreme Court to expressly rely on deep theories of moral wrongfulness or criminalization, since this would embroil the judiciary in deeply contentious political questions it lacks the legitimacy or expertise to authoritatively settle. Better, the Sunsteinian might say, simply to gesture towards various courses of action that people of very different political orientations can agree are inappropriate targets of criminalization, leaving the basis for that conclusion unarticulated.[99] Such an approach would be consistent with the Supreme Court’s tendency not to wed itself to any particular theory of criminalization.[100]

The Sunsteinian lens, however, seems[101] to treat the absence of abstract moral theorizing as itself a political strategy on the part of the courts—a means of avoiding the controversy and censure which would (surely) result if judges overtly imposed their own theory of morality or criminalization upon the legislature by articulating the practical implications of that theory on a piecemeal basis. I want to propose a slightly different explanation; one that treats the decision not to articulate a deep moral or political theory as principled rather than purely pragmatic. On this view, courts should take into account the widespread acceptance or toleration of certain practices and courses of action, not because they could necessarily be defended if subjected to deep moral scrutiny, but because they are widely accepted or tolerated.

C. Fuller on the Significance of Interactional Expectancies

To make sense of this idea, we need to keep in mind that rules can have normative force—that is, they can function as rules[102]—without having been enacted (or “made”) by an authoritative person or institution. They may, instead, draw their normative force from the sheer fact that a “stable set of interactional expectancies” or “intermeshing anticipations” has built up over time among the actors themselves in a given social situation.[103] Such “implicit rules” are not consciously created; on the contrary, they are typically by-products of sustained patterns of interaction which have given rise to mutual expectations.[104] It is the very fact that these mutual expectations already exist—and are known to exist by the actors themselvesthat gives them a reason to act in accordance with those expectations.[105] As Postema puts it: “[I]mplicit rules emerge as focal points around which persons who must coordinate[106] their actions form reliable expectations of others knowing that their counterparts are doing the same thing.”[107]

Fuller took the view that many of these implicit rules, conventions, customs, informal practices, and usages deserve to be characterized as law precisely because, like enacted law, they structure interactional expectancies and therefore guide actors. On several occasions, Fuller lamented the lack of attention paid by legal philosophers and scholars to implicit law.[108] This is not just because we can learn something about the nature of law generally by bringing unwritten norms and customs into the frame.[109] Fuller contended that it is impossible to understand enacted law in modern industrialized societies without referring to implicit law.[110] For example, contracts create mutual expectations between parties, but the practical significance of the language can only be appreciated in light of unwritten commercial norms to which the parties are also expected to conform.[111] Similarly, the Constitution creates mutual expectations for those in government, but again the written language can only be adequately understood if one reads it in light of a vast range of unwritten conventions and structural norms.[112]

For my purposes though, I am most interested in the suggestion that one cannot effectively interpret legislation without appreciating the unwritten customs and practices that structure the expectations of those the legislator intends to guide.[113] If a norm or practice exists in the first place it is because participants expect one another to treat it as a reason for action, and each knows that the others have this expectation. In the absence of a clear legislative signal that the norm or practice is to cease, participants may hesitate to read a statute in that way, anticipating that others will assess the significance of their conduct according to the pre-existing norm and expect them to do so as well.[114] Knowing this, a legislature that indeed seeks to dislodge or displace that norm or practice will provide a clear statement[115] to that effect, given that the provision at issue may otherwise fail to discharge its guidance function. Postema remarks:

[W]e expect that citizens’ understanding of what the law requires of them will determine at least in part their decisions and actions. But that understanding depends on their expectation of how lawmaking ... officials are likely to understand it. Similarly, officials authorized to enact ... the laws must anticipate how citizens will take up the laws they make ... , that is, how the rules are likely to figure in the practical reasoning of citizens. Otherwise, they will not be able to direct or guide actions in such a way as to achieve the substantive aims of the law.[116]

One might ask how citizens’ interactional expectations of one another could be displaced by criminal offences—specifically, those offences that courts have interpreted using the presumption of restraint. Unfortunately, Fuller’s own analysis of the role of implicit law in the criminal sphere is not crystal clear. I would note, however, that he took a quite broad view of “human interaction.” He argued, for example, that offenses against deities and spirits—though we might instinctively regard these as targeting “private” conduct—affect interactional expectancies insofar as they go to the “significance one man’s acts may have for his fellows.”[117] Fuller observed, moreover, that certain customs and rituals go to interactional expectancies insofar as they affect how people in the community understand and make sense of their relationship with one another.[118]

Fuller’s comments about the criminal law are also instructive. He acknowledged that many people would not think it obvious that the offence of murder facilitates interaction. But, he argued, it does regulate how people think it appropriate to respond to killings: they do not, Fuller observed, think it fit to engage in blood feuds.[119] Furthermore, the laws governing self-defence and arrest guide individuals as they consider whether and when it is appropriate to use force against others.[120] Here too Fuller emphasized that interactional expectancies are engaged insofar as a norm—in this instance, enacted law—guides citizens’ appreciation of the significance of an act or status, such that they can appropriately respond to the actor and others now and in the future.

Admittedly, Fuller suggested that so-called “victimless crimes” do not facilitate human interaction, remarking that they serve to prevent interaction instead.[121] But he did not press the point, and it is difficult to see how such a position could be correct given his earlier observations that norms facilitate human interaction so long as they shed light on the social significance of one another’s conduct and status.[122] After all, the prohibition on certain transactions surely sends a message about the social significance of certain goods and activities—for example, the treatment of blood or sex as an intrinsic rather than instrumental good.[123] It also conveys that invitations to engage in certain transactions should be resisted, and arguably stigmatizes those who engage in them.[124] Relatedly, it signals what, if any, significance should be given to the apparent “consent” of the parties, and how we may respond when it is given.[125]

Taking all this on board, the case authorities pertaining to the presumption of restraint clearly involve norms pertaining to human interaction (in the broad Fullerian sense).[126] The pedestrian knows not to admonish Justice Cory’s window-shopper or leafletting politician for “loitering,” who in turn know they are free to carry on as they are.[127] The friends of government employees know that, when invited out for coffee, they are entitled to accept, and the employees themselves know that their friends will not see such invitations as potential traps.[128] Churchgoers appreciate that they may chastise Justice Dickson’s hat-wearing boor, but not forcibly expel him or alert the authorities.[129] The onlooker who sees a parent wrapping a scarf around the neck of her struggling child knows that she should not intervene, and the parent knows that she is entitled to continue.[130] More controversially, the victim of seductive lying knows that she may complain about the deception, but also knows that she is not (at least, for that reason alone) the victim of an “assault.”[131]

In each instance, an expansive interpretation of the respective offences could radically upset or shift social understandings of the conduct in question, and what responses would be expected or regarded as eligible.[132] To the extent that Parliament seeks to unsettle people’s customary ways of treating and reacting to conduct regarded as benign or only trivially wrongful, Fuller’s analysis suggests that it must send reasonably clear signals to that effect. If it does not, it will fail to provide the desired guidance. Insofar as the presumption of restraint encourages Parliament to send such signals, the presumption can be regarded as a means of ensuring that Parliament can function more effectively as a lawmaker.[133]

For the sake of clarity, it is worth reiterating that nothing in this account presupposes the correctness or even reasonableness of the unwritten norm in question. Patterns of interaction often have a rational foundation, in the sense that they emerged out of a felt need to resolve some sort of coordination problem.[134] We need not assume, however, that they represent the best (or even second- or third-best) solution to that dilemma. If one set out to draft a blueprint for the just society from scratch, one might well conclude that different rules would be preferable. Unwritten norms are sustained by the widely held perception that they are reasonable,[135] but that perception can itself be a function of the fact that, once a stable norm is in place, it exerts a normative pull on actors, and can even assume an aura of naturalness or inevitability.[136] This may be the very reason why legislation is needed; that is, to sweep away pernicious conventions and practices and replace them with a more just normative regime.[137] The presumption of restraint, on my reading, does not preclude or block that sort of legislative intervention. It only recognizes that, in the absence of some clear statement, it may be ineffective—and that Parliament could not have intended that.

I would point out that grounding the presumption of restraint in the criminal law’s internal morality finds oblique support in other aspects of Fuller’s theory. In The Morality of Law, Fuller suggests that the guidance function of law may be impaired where enacted law is changed so suddenly and frequently that subjects are unable to orient their actions by them.[138] The clear suggestion is that, in trying to promulgate one norm, one cannot (as a prudential matter) be indifferent to the norms one seeks to displace. Moreover, one can also see a link to Fuller’s criterion of comprehensibility.[139] The legislator who wishes to guide her subjects must do so in a language they can understand. But the fact that the legislator “speaks” in a common tongue is neither here nor there if her subjects are so accustomed to thinking of a course of action as permissible and proper that they do not recognize her prohibition as a prohibition on that.


I would acknowledge that real concerns arise with this account. Chiefly, we may wonder if the courts are qualified to say whether a practice or course of action is widely regarded as benign or only trivially wrongful. The Supreme Court has not cited any evidence to support the premise that this or that course of action satisfies such a test, nor indicated that any is necessary. Furthermore, we may wonder if it matters that a course of action is regarded as benign or insufficiently wrongful by one community or region rather than another. Should the presumption apply in such instances? It is perhaps significant that in more recent cases like Mabior and Hutchinson the Court has not referred to concrete examples of “obviously” benign or insufficiently wrongful conduct, preferring to state merely that the criminal law should be reserved for “reprehensible” conduct. This rhetorical choice may reflect the fact that, in both cases, the concrete implications of an expansive reading were obvious. It may, however, reflect a reluctance within the Court to confront these questions head-on.

I have set out to offer one possible rationale for a persistent interpretive strategy employed by the courts in criminal cases. In particular, I have suggested that, by looking to the idea of an “internal morality of criminal law,” we may find a way to reconcile the presumption of restraint with the Supreme Court’s broadly deferential attitude to Parliament’s use of the criminal law power. Even if this proposed solution does not succeed, though, I hope to have shown that the presumption generates very real questions deserving greater attention.