[1] Mr. Thomas Chan and Mr. David Sullivan share similar, tragic experiences. In separate incidents, while in the throes of drug-induced psychoses and without any discernible motive, both men attacked and stabbed loved ones. Mr. Chan, who became intoxicated after consuming "magic mushrooms", killed his father and grievously injured his father's partner. Mr. Sullivan, who had become intoxicated after consuming a heavy dose of a prescription drug in a suicide attempt, repeatedly stabbed his elderly mother. Both men allege that they were in a state of automatism at the time of the attacks.

[2] Automatism is defined as "a state of impaired consciousness, rather than unconsciousness, in which an individual, though capable of action, has no voluntary control over that action": R. v. Stone, 1999 CanLII 688 (SCC), [1999] 2 S.C.R. 290, [1999] S.C.J. No. 27, at para. 156, per Bastarache J. Involuntariness is therefore the essence of automatism. The "mind does not go with what is being done": R. v. Rabey, 1980 CanLII 44 (SCC), [1980] 2 S.C.R. 513, [1980] S.C.J. No. 88, at p. 518 S.C.R., citing R. v. K. (1970), 1970 CanLII 431 (ON SC), [1971] 2 O.R. 401, [1970] O.J. No. 1761 (H.C.J.), at p. 401 O.R.

[3] Persons in a state of automatism may have the benefit of a "defence" when they engage in otherwise criminal conduct, even though automatism is not a justification or excuse: R. v. Luedecke (2008), 93 O.R. (3d) 89, [2008] O.J. No. 4049, 2008 ONCA 716, at [page361] para. 56. Instead, automatism is treated as negating the crime. It is referred to as a defence because the accused bears the burden of establishing automatism. In Luedecke, at para. 56, Doherty J.A. explained the underlying principles:

[4] There are two branches to the defence of automatism. The mental disorder defence, codified in s. 16 of the Criminal Code, R.S.C., 1985, c. C-46, applies where involuntariness is caused by a disease of the mind, since those who are in a state of automatism are incapable of appreciating the nature and quality of their acts or of knowing at the time of their conduct that it is morally wrong ["mental disorder automatism"]. If successful, a mental disorder automatism defence will result in a not criminally responsible verdict, with the likelihood of detention or extensive community supervision.

[5] The alternative branch, the common law automatism defence, applies where the involuntariness is not caused by a disease of the mind ["non-mental disorder automatism"]. Where a non-mental disorder automatism defence succeeds, the accused is acquitted.

[6] Mr. Chan and Mr. Sullivan each relied on non-mental disorder automatism as their primary defence. The hurdle they each faced is that their non-mental disorder automatism claims arose from their intoxication, and each man was charged with violent offences. Yet, s. 33.1 of the Criminal Code ["s. 33.1"] removes non-mental disorder automatism as a defence where the state of automatism is self-induced by voluntary intoxication and the offence charged includes "as an element an assault or any other interference or threat of interference by a person with the bodily integrity of another person" [a "violence-based offence"].

[7] Mr. Chan tried to overcome the impediment s. 33.1 presented to his non-mental disorder automatism defence by applying to have the section declared to be of no force or effect under s. 52(1) of the Constitution Act, 1982, Schedule B to the Canada Act, 1982 (U.K.), 1982, c. 11, as contrary to the Canadian Charter of Rights and Freedoms. The trial judge agreed with Mr. Chan that s. 33.1 is in prima facie violation of ss. 7 and 11(d) of the Charter but upheld the constitutionality of s. 33.1 under s. 1 of the Charter, as a demonstrably justifiable limit on the Charter rights Mr. Chan invoked. [page362]

[8] Mr. Sullivan argued that s. 33.1 did not prevent him from relying on the non-mental disorder automatism defence because his intoxication was not voluntary, having resulted from a suicide attempt. The trial judge rejected this contention and found s. 33.1 to apply.

[9] Both Mr. Chan and Mr. Sullivan defended themselves, in the alternative, by claiming that if they were not experiencing non-mental disorder automatism at the time of their respective attacks, they were experiencing mental disorder automatism. Neither Mr. Chan nor Mr. Sullivan's mental disorder defences succeeded, and both men were convicted of the violence-based charges they faced.

[10] They now appeal. They both raise additional grounds of appeal, but their appeals have in common that they both challenge convictions claiming that s. 33.1 unconstitutionally deprived them of access to the non-mental disorder automatism defence. Mr. Chan does so by challenging the trial judge's rulings. Mr. Sullivan raises the constitutional validity of s. 33.1 for the first time on appeal, as his case is still in the system. The Crown concedes that if Mr. Chan's s. 33.1 challenge succeeds, Mr. Sullivan would also be entitled to the benefit of that ruling. We therefore heard their appeals together and I address them together in this joint decision.

[11] For the reasons below, I would find that s. 33.1 is indeed unconstitutional and I would declare it to be of no force or effect.

[12] Since Mr. Chan was convicted only of offences that include an element of assault, and those convictions depended upon s. 33.1, I would allow his appeal in its entirety and order a new trial.

[13] Mr. Sullivan's violence-based convictions must also be set aside, for the same reason. The Crown agrees that, in these circumstances, verdicts of acquittal should be substituted for Mr. Sullivan's violence-based convictions, and I would do so. Mr. Sullivan also appeals four breach of recognizance convictions for contacting his sister while in custody, contrary to a non-communication order. As I will explain, I would reject his appeal of these convictions.

[14] I will begin with Mr. Chan's appeal, since this is where the bulk of the arguments relating to the constitutional validity of s. 33.1 were made.

[15] The consumption of "magic mushrooms", containing the active ingredient psilocybin, triggered Mr. Chan's extreme intoxication. After an evening watching a hockey game at a pub, Mr. Chan, a high school student, and three friends ingested magic [page363] mushrooms in the basement of Mr. Chan's mother's home, where he lived. Mr. Chan had used magic mushrooms before, and the experience had always been pleasant and uneventful. Within a half-hour of consuming the magic mushrooms, his friends were high, but Mr. Chan was not. He took an additional quantity of mushrooms.

[16] Forensic toxicologist, Dr. Daryl Mayers, testified that psilocybin is a "pretty safe" drug in terms of toxicity. On this occasion, it was anything but. A few hours after ingesting the drugs, Mr. Chan's behaviour changed. He expressed that he was scared, began speaking in gibberish, and at some point, ran upstairs to his mother's room where she and her boyfriend, Mr. Jeff Phillips, were sleeping. Mr. Chan turned on the lights and refused to turn them off. He began calling his mother and sister "Satan" and "the Devil" and claimed to "see the light". Mr. Chan ran outside, where it was below freezing and snowing, wearing only a pair of pants.

[17] Mr. Chan then ran to his father's [Dr. Chan's] house, which was just around the corner. Outside of his father's house, he tried to fight with one of his friends who had followed him, and he smashed a car window. Several neighbours reported that they heard a male voice yelling phrases such as, "This is God's will" and "I am God".

[18] Mr. Chan broke into his father's house through a window even though he customarily gained entry by using finger-print recognition on the home security system. He confronted Dr. Chan in the kitchen. Dr. Chan said, "Thomas, it's Daddy. It's Daddy", but Mr. Chan did not appear to recognize him and stabbed him repeatedly. Dr. Chan died of his injuries.

[19] Mr. Chan then began attacking Dr. Chan's partner, Ms. Lynn Witteveen. Ms. Witteveen said, "Thomas, it's Lynnie, it's Lynnie. I love you", but she did not think he recognized her. Mr. Chan stabbed her in the abdomen, arm, back, and chest. At some point after Ms. Witteveen called 9-1-1, Mr. Chan also stabbed her right eye and slashed her neck.

[20] When the police arrived, Mr. Chan immediately complied with their demands to raise his hands and drop to the ground. After the police handcuffed him, he began to struggle. Police Constable Heenan described him as having "super-strength".

[21] Mr. Chan offered alternative arguments to support his claim that s. 33.1 is unconstitutional. First, he urged that since s. 33.1 was declared to be unconstitutional in R. v. Dunn, 1999 CanLII 36525 (ON SC), [1999] O.J. No. 5452, 28 C.R. (5th) 295 (S.C.J.), it no longer had force or effect in Ontario, and that the trial judge was bound to disregard it. In the alternative, he asked the trial judge to find that the provision [page364] violates ss. 7 and 11(d) of the Charter and cannot be demonstrably justified under s. 1.

[22] As described, the trial judge denied Mr. Chan's Charter challenge. Since s. 33.1 applied, Mr. Chan's non-mental disorder automatism defence was unavailable.

[23] Mr. Chan's mental disorder defence was also denied. The trial judge found that although Mr. Chan's rugby career had left him with cognitive deficits linked to a mild traumatic brain injury, and although Mr. Chan was incapable at the time of the attack of knowing that his actions were morally wrong, his psychosis was the direct result of self-induced intoxication through the ingestion of psilocybin. Since the psychosis was not caused by a disease of the mind, the mental disorder defence would not apply.

[24] In his reasons for judgment in finding Mr. Chan guilty of the grave charges he faced, the trial judge remarked [at para. 141], "Mr. Chan is not a danger to the public. He is a good kid who got super high and did horrific things while experiencing a drug-induced psychosis."

[25] Mr. Chan appeals his convictions. He argues that the trial judge erred in denying his Charter challenge to s. 33.1, both because the trial judge was bound by the declaration of unconstitutionality in Dunn, and that, in any event, s. 33.1 cannot be demonstrably justified under s. 1 of the Charter. He asks that acquittals be entered if either of these grounds of appeal succeed.

[26] Alternatively, Mr. Chan argues that the trial judge erred in rejecting the mental disorder defence and asks us to set aside his convictions and to either substitute findings of not criminally responsible by reason of mental disorder or order a new trial.

[27] The Crown contends that the trial judge was correct in the ultimate conclusions he reached but erred in finding s. 33.1 to be in prima facie violation of the Charter.

[28] Mr. Chan's appeal therefore raises the following issues:

[29] I agree with the trial judge that he was not bound by prior authority to treat s. 33.1 as having no force or effect. I also agree with the trial judge that s. 33.1 violates ss. 7 and 11(d) of the Charter. However, the trial judge erred in finding that these violations are demonstrably justifiable under s. 1. Mr. Chan's appeal must be allowed.

[30] Mr. Chan asks us to substitute verdicts of acquittal. I would not do so and would order a new trial. Given this, I need not determine whether the trial judge erred in rejecting the mental disorder defence. This ground of appeal is largely fact driven, and if it arises again, that issue should be decided by the trial judge at the re-trial.

[31] Mr. Chan argues that once a superior court judge declares a law to be unconstitutional, that declaration is binding on other superior court judges, unless the Crown has successfully appealed that decision. He recognises that this position is inconsistent with the ordinary principles of stare decisis, which hold that lower courts are required to follow only binding precedent of higher courts but are not strictly bound to follow earlier decisions in the same court: Robert J. Sharpe, Good Judgment: Making Judicial Decisions, (Toronto: University of Toronto Press, Scholarly Publishing Division, 2018), at pp. 153-155.

[32] Mr. Chan points out that it in Nova Scotia (Workers' Compensation Board) v. Martin, [2003] 2 S.C.R. 504, [2003] S.C.J. No. 54, 2003 SCC 54, at para. 28, the Supreme Court of Canada recognized that "[t]he invalidity of a legislative provision inconsistent with the Charter does not arise from the fact of its being declared unconstitutional by a court, but from the operation of s. 52(1)". The result, he says, is that s. 52(1) governs the binding effect of superior court declarations of unconstitutionality, and the principles of stare decisis do not. Section 52(1) provides that "any law that is inconsistent with the provisions of the Constitution is, to the extent of the inconsistency, of no force or effect". Therefore, once a superior court judge makes a s. 52(1) declaration, all other superior court judges within the province are bound to treat it as such. On that basis, he contends that since s. 33.1 was found to be unconstitutional in Dunn prior to Mr. Chan's prosecution, and that decision was not appealed and set aside, the trial judge erred in relying on s. 33.1, as it was of no force or effect. [page366]

[33] As the decision in R. v. McCaw, [2018] O.J. No. 4134, 2018 ONSC 3464, 48 C.R. (7th) 359 (S.C.J.) reveals, superior court case law in Ontario is split on whether this is correct. There does not appear to be appellate authority directly on point, although in an obiter comment made in another context, in R. v. Boutilier, [2016] B.C.J. No. 71, 2016 BCCA 24, 332 C.C.C. (3d) 315, at para. 45, Neilson J.A. commented that a declaration is "a final order in the proceeding directed at the constitutionality of [the impugned provision], binding on the Crown and on other trial courts of [the] province" (emphasis added).

[34] With respect, I cannot agree. I am persuaded that the ordinary principles of stare decisis apply, and that the trial judge was not bound by the Dunn decision. The authorities relied upon by Mr. Chan do not purport to oust these principles. In Nova Scotia (Workers' Compensation Board), at para. 28, Gonthier J. was simply explaining that a provision that is inconsistent with the Constitution "is invalid from the moment it is enacted, and a judicial declaration to this effect is but one remedy amongst others to protect those whom it adversely affects". He was not attempting to alter the principles of stare decisis where s. 52(1) declarations have been made.

[35] Similarly, in none of the other passages relied upon by Mr. Chan was the Supreme Court of Canada purporting to oust the principles of stare decisis where s. 52(1) declarations have been made. The passages he refers to proclaim that after a s. 52(1) declaration is made, the law: is invalid "for all future cases"; "cannot be enforced"; and is "null and void, and is effectively removed from the statute books", such that "[t] he ball is thrown back into Parliament's court": see respectively Nova Scotia (Workers' Compensation Board), at para. 31; Canada (Attorney-General) v. Hislop, [2007] 1 S.C.R. 429, [2007] S.C.J. No. 10, 2007 SCC 10, at para. 82; and R. v. Ferguson, [2008] 1 S.C.R. 96, [2008] S.C.J. No. 6, 2008 SCC 6, at para. 65. These passages describe the effects of a s. 52(1) declaration that has been affirmed or made by the Supreme Court of Canada, the apex court. Those passages cannot be taken as describing the effect of declarations made by lower courts. After all, declarations made by trial courts are subject to appeal, and if overturned on appeal, will have no effect. Even on Mr. Chan's theory, superior court declarations are not binding outside of the province in which they are made. In these circumstances, it cannot be said that a superior court declaration determines the validity or enforcement of the statute "for all future cases", effectively removes the impugned provision from the statute books, or throws the ball back into Parliament's court. These things happen only if the Supreme Court of Canada affirms or makes a s. 52(1) declaration. [page367]

[36] Nor can Mr. Chan find assistance in McLachlin C.J.C.'s observation in Canada (Attorney General) v. Bedford, [2013] 2 S.C.R. 1101, [2013] S.C.J. No. 72, 2013 SCC 72, at paras. 43-44, that "the common law principle of stare decisis is subordinate to the Constitution". Bedford recognized that the principles of stare decisis cannot be relied upon to perpetuate a statute, where that statute is unconstitutional when viewed again through a new lens. Specifically, a trial judge can depart from binding precedent when "a new legal issue is raised, or if there is a significant change in the circumstances or evidence": Bedford, at para. 44. However, Bedford does not hold that the principles of stare decisis are ousted whenever constitutional issues are at stake.

[37] There is good reason why not. Whereas Bedford compromises stare decisis to promote accurate constitutional outcomes, the compromise on stare decisis proposed by Mr. Chan has the potential to discourage accuracy. For example, three superior court judges in succession could find a provision to be constitutional, but the fourth judge's ruling to the contrary would be the only one to have full force or effect in the province. Unless that fourth decision is appealed, it becomes the law in the province. The Crown can no longer rely on the provision; therefore, decreasing the prospect that the issue of constitutional validity would make it before the provincial appellate court. The development of the law would be driven by coincidence in the sequence of trial level decisions and the fortuity of discretionary decisions about whether to appeal, when it should be determined by the quality of the judicial ruling.

[38] The application of the principles of stare decisis to s. 52(1) declarations made by superior court judges does not mean that a superior court declaration will have no effect in other cases. Other superior court judges should respect an earlier declaration of unconstitutionality, absent cogent reason to conclude that the earlier declaration is plainly the result of a wrong decision: R. v. Scarlett, [2013] O.J. No. 644, 2013 ONSC 562, at para. 43; Hansard Spruce Mils Ltd. (Re), 1954 CanLII 253 (BC SC), [1954] B.C.J. No. 136, [1954] 4 D.L.R. 590 (S.C.), at p. 592 D.L.R. It is obvious that a superior court judge cannot determine that there is cogent reason to conclude that the earlier decision is plainly wrong without the benefit of argument, facilitated by fair notice to the parties. Therefore, where a party seeks to rely on a statutory provision that has been declared to be unconstitutional by a superior court judge, a subsequent trial judge should apply that earlier declaration of invalidity and treat the statutory provision as having no force or effect, unless the underlying constitutional issue has been raised by the Crown before them through submissions that the earlier decision is plainly wrong. In this way, the principles of stare decisis can operate, while recognizing that [page368] the effect of a s. 52(1) declaration is not confined to the litigation in which the declaration is made.

[39] The application of the ordinary principles of stare decisis to s. 52(1) declarations in no way collapses the structural distinction between s. 52(1) and s. 24(1) of the Charter, or amounts to a constitutional exemption, as Mr. Chan argues. The fact that cases at the superior court trial level may produce different outcomes for respective accused persons does not mean that the remedies are personal. The disparity in outcome simply reflects the developing state of the authority on the constitutional validity of a provision, as advanced by judges of competent jurisdiction.

[40] The trial judge was correct in finding that he was not bound by Dunn. He was also correct in considering the issue anew, as the issue of the constitutionality of s. 33.1 was put before him, and the authority he encountered was inconsistent. He had no choice, in the circumstances, but to consider whether to deviate from Dunn.

[41] I would therefore reject this ground of appeal.

[42] The trial judge was correct in finding s. 33.1 to be in prima facie violation of both ss. 7 and 11(d) of the Charter. Section 33.1 violates each of the constitutional principles that were identified by Cory J. for the majority in R. v. Daviault, 1994 CanLII 61 (SCC), [1994] 3 S.C.R. 63, [1994] S.C.J. No. 77. In Daviault, the Supreme Court of Canada modified the common law rule that eliminated the defence of extreme intoxication because the common law rule was in breach of the Charter in three ways. I will describe these breaches as "the voluntariness breach", "the improper substitution breach", and "the mens rea breach". Although there has been some variation in articulation and emphasis, virtually all the judges who have considered this issue have found that the legislation breaches the Charter in one or more of these respects.

[43] I will begin by introducing the relevant constitutional principles in the context of the Daviault decision. I will then address and reject general arguments made before us that the constitutional principles recognized in Daviault do not govern whether s. 33.1 is in prima facie violation of the Charter. I will then analyse these principles in detail and explain why s. 33.1 contravenes ss. 7 and 11(d) of the Charter in these respects.

[44] Mr. Daviault was charged with sexual assault. The sexual act he was charged with committing occurred after Mr. Daviault had been drinking heavily. He claimed he was so extremely [page369] intoxicated that, at the time of the act, he was in a state of automatism. The decision in R. v. Leary (1977), 1977 CanLII 2 (SCC), [1978] 1 S.C.R. 29, [1977] S.C.J. No. 39, imposed an impediment to Mr. Daviault's attempt to rely on his extreme intoxication as a defence. Under the "Leary rules" voluntary intoxication can be presented as a defence only to a "specific intent offence", but not a "general intent offence", and sexual assault is a general intent offence.

[45] There are policy reasons that support criminal consequences when general intent offences are committed by those who choose to become intoxicated. However, the primary distinction between general intent and specific intent offences lies in the complexity of the thought and reasoning process required to commit the relevant offence: R. v. Tatton, [2015] 2 S.C.R. 574, [2015] S.C.J. No. 33, 2015 SCC 33, at para. 21. The mental states required to commit general intent offences simply relate to the performance of the illegal act with no further ulterior purpose; therefore, they involve "minimal mental acuity": Tatton, at paras. 35, 41; Daviault, at p. 89, per Cory J.; and R. v. Bernard, 1994 CanLII 61 (SCC), [1988] 2 S.C.R. 833, [1994] S.C.J. No. 77, at p. 863 S.C.R., per McIntyre J. These mental states tend to be so basic or rudimentary that, ordinarily, it is not realistic to believe that intoxication could prevent an accused who has committed the prohibited act from having the mental state the offence requires. Proof of intoxication is typically irrelevant in general intent offences, as the requisite mental state can usually be inferred from the mere commission of the act: Bernard, at p. 878 S.C.R., per McIntyre J.; Daviault, at p. 123 S.C.R., per Sopinka J. (dissenting on other grounds).

[46] In contrast, specific intent offences tend to require more complex mental elements over and above the minimal intent required for general intent offences: Daviault, at pp. 123-24 S.C.R., per Sopinka J. (dissenting on other grounds). The actus reus must be coupled with an intent or purpose going beyond the mere performance of the prohibited act; for example, striking a blow with the intent to kill: Bernard, at p. 863 S.C.R., per McIntyre J. For specific intent offences, the fact of intoxication may well be relevant in deciding whether the accused had the more complex specific intent and so, proof of self-induced intoxication is permitted.

[47] The majority in Daviault qualified this sharp general intent offence/specific intent offence divide by accepting the view expressed by Wilson J. in Bernard, at p. 887 S.C.R., that evidence of extreme intoxication involving an absence of awareness akin to a state of automatism is not irrelevant on issues of general intention. Such evidence can raise a reasonable doubt as to the existence of even the minimal intent required for a general intent offence such as sexual assault. Given that this is so, the Daviault [page370] majority concluded that the Charter requires both the admissibility of evidence of extreme self-induced intoxication, as well as access to the defence of automatism, even when the automatism is the result of self-induced intoxication. To do otherwise, would contravene the right to life, liberty, and security of person in a manner that does not accord with the principles of fundamental justice (Charter, s. 7) and the presumption of innocence (Charter, s. 11(d)). More specifically:

[48] The Daviault majority went on to find that the identified Charter violations could not be justified under s. 1 of the Charter. It held that there is no pressing and substantial purpose in preventing access to the "rare and limited defence" of automatism arising from self-induced intoxication, and the deleterious effects of doing so are not overcome by proportionate benefits: Daviault, at p. 103 S.C.R.

[49] The Daviault majority did hold, at p. 101 S.C.R., however, that it is a reasonable limitation on the Charter rights identified to require accused persons to establish automatism with the assistance of expert evidence, on the balance of probabilities.

[50] The Daviault decision, with its notion that extreme intoxication could provide a pathway to exoneration for sexual [page371] assault, created significant public outcry. Parliament responded by passing Bill C-72, An Act to amend the Criminal Code (self-induced intoxication), 1st Sess, 35th Parl, 1995 (assented to 13 July 1995), S.C. 1995, c. 32, which added s. 33.1 to the Criminal Code. Bill C-72 included an extensive preamble [the "Preamble"]:

[51] Section 33.1 of the Criminal Code provides:

[52] Arguments were presented before us that the principles identified in Daviault do not govern the constitutional validity of s. 33.1. I do not agree.

[53] First, I do not accept the Crown's contention that the Charter principles identified in Daviault apply only to common law rules, not statutory ones such as s. 33.1, or that Daviault "provides only that courts cannot water down the requirements of statutory offences by omitting the need for statutorily defined essential elements" (emphasis in original). The Daviault decision is not about the division of powers between Parliament and the courts. The sole reason that the Supreme Court of Canada reconfigured the common law Leary rules in Daviault was that, without reconfiguration, the Leary rules infringed principles of fundamental justice assured by s. 7 of the Charter, as well as the presumption of innocence under s. 11(d). Those principles of fundamental justice were not created in Daviault. They had already been recognized by other Supreme Court of Canada authority. Nor does the reach or definition of those Charter principles vary depending upon whether the law being tested is a common law or statutory rule. As s. 52(1) of the Constitution Act, 1982 makes clear, subject to s. 1 of the Charter, these constitutionally-protected principles must be respected by "any law", common law or statutory. If the law does not do so, it will be of no force or effect to the extent of the inconsistency.

[54] In Daviault, the Supreme Court of Canada occasionally referenced the limits on "judicially developed policy" or the ability of courts to eliminate elements of a crime. These are contextual allusions to the fact that the rules under challenge in that case were common law rules. In making these comments, the Daviault majority was not attempting to confine the reach of the constitutional principles relied upon. When Cory J. recognized that it was open to Parliament to legislate in this area, he was accepting that there are ways for Parliament to address extreme intoxication, but he was [page373] not suggesting that Parliament could do so in disregard of the constitutional principles described.

[55] To be clear, no one questions that Parliament has the authority to amend criminal offences, and that courts do not. The instant point is that when Parliament purports to make statutory changes, it must do so consistently with the Charter, and in determining whether this is so, the Charter principles identified in Daviault apply.

[56] Nor do I accept the argument advanced by the intervener, the Women's Legal Education Action Fund ("LEAF"), that s. 7 of the Charter requires "internal balancing" in identifying the relevant principles of fundamental justice for consideration. Specifically, LEAF argued that in determining whether there is a prima facie breach, we must balance the accused's interests and public interests, such as equality and the human dignity of women and children, who are disproportionally victimized by intoxicated offenders.

[57] Generally, there is no place for internal balancing in defining the principles of fundamental justice. As Lamer C.J.C. explained in R v. Swain, 1991 CanLII 104 (SCC), [1991] 1 S.C.R. 933, [1991] S.C.J. No. 32, at p. 937 S.C.R., it is not appropriate to thwart the exercise of the accused's s. 7 rights by trying to bring societal interests into the principles of fundamental justice to limit those rights. If societal interests should limit those rights, it is for the Crown to show this under s. 1. This was the law when Daviault was decided and it remains the law, having recently been reaffirmed in Bedford, at paras. 124-27, and Carter v. Canada (Attorney General), [2015] 1 S.C.R. 331, [2015] S.C.J. No. 5, 2015 SCC 5, at paras. 78-80.

[58] I recognize that in R. v. Mills, 1999 CanLII 637 (SCC), [1999] 3 S.C.R. 668, [1999] S.C.J. No. 68, the Supreme Court of Canada did conduct internal balancing of competing Charter-protected interests. This exercise was required because the issue in Mills was whether the legislative accommodation between the privacy and equality rights of sexual offence complainants, on the one hand, and the right of the accused to access information, on the other, infringed Mr. Mill's s. 7 right to full answer and defence. No such internal balancing is required in this case. It is not about the constitutionality of a legislated compromise between protected interests. Moreover, as about to be explained, the reach of the principles of fundamental justice at issue have already been authoritatively determined, and this has occurred in a body of law that has not engaged in internal balancing. I propose to rely on these principles and to consider the important interest identified by LEAF under s. 1.

[59] Finally, at trial, the Crown argued that "the court must follow the analysis in Bedford, meaning that the court must measure s. 33.1 against the principles of "arbitrariness, [page374] overbreadth and gross disproportionality": see R. v. Chan, [2018] O.J. No. 4731, 2018 ONSC 3849, 365 C.C.C. (3d) 376, at para. 92. None of the parties before us argued, as the trial Crown had, that Bedford has changed the way that s. 7 analysis is to be conducted, but the impact of Bedford and Carter was raised during oral argument and by my colleague in his concurring decision. I will therefore address the issue briefly.

[60] I do not share my colleague's view that we are bound by Bedford or Carter to apply the principles of "arbitrariness, overbreadth and gross disproportionality" to the issue of whether s. 33.1 limits s. 7 Charter rights. Arbitrariness, overbreadth and gross disproportionality are engaged if the s. 7 challenge is that the effect of the law is not connected to its objective ("arbitrariness"), that the law catches situations that have no connection to its objective ("overbreadth"), or that the law imposes consequences that are grossly disproportionate to its objective ("gross disproportionality"): Bedford, at paras. 97-105. These principles all stem from what Professor Hamish Stewart calls "failures of instrumental rationality": Fundamental Justice: Section 7 of the Canadian Charter of Rights and Freedoms (Toronto: Irwin Law, 2012), at p. 151, cited in Bedford, at para. 107. Many principles of fundamental justice have little or nothing to do with "instrumental rationality", in any of these senses.

[61] The principles of fundamental justice identified in Daviault, which the appellants now rely upon, do not address the link between the objective and effects of s. 33.1. Instead, these principles identify what is constitutionally required before a criminal conviction is permitted. In other words, they impose constitutional limits on criminal accountability. The principles affirmed in Daviault have not been modified in any way by subsequent authority, as the trial Crown suggested, nor do I read the Bedford and Carter cases as requiring consideration of arbitrariness, overbreadth and gross disproportionality in all s. 7 cases, as my colleague maintains. Those principles were considered in Bedford and Carter because, in substance, the challenges before those courts alleged failures of instrumental rationality: see Bedford, at para. 96; Carter, at para. 46. They were not considered in R. v. Morrison, [2019] 2 S.C.R. 3, [2019] S.C.J. No. 15, 2019 SCC 15, 432 D.L.R. (4th) 637, at paras. 74-91, where, as in this case, the challenge was to the compliance of a Criminal Code provision with the minimum level of constitutionally required fault. In my view, the trial judge was correct in rejecting the trial Crown's invitation to consider these principles, and in addressing only the principles identified in Daviault. I will proceed in the same manner. [page375]

[62] I do not accept that the Supreme Court of Canada implicitly suggested that s. 33.1 is constitutionally valid because it did not cast doubt on the constitutionality of s. 33.1 when deciding R. v. Bouchard-Lebrun, [2011] 3 S.C.R. 575, [2011] S.C.J. No. 58, 2011 SCC 58. LeBel J. noted expressly that the constitutional validity of s. 33.1 was not before the court in that case: at para. 28. The issue was whether the trial judge erred by treating s. 33.1 as limiting the scope of the mental disorder defence in s. 16 of the Criminal Code, an entirely different question.¹ The trial judge was correct in rejecting the suggestion that Bouchard-Lebrun supports the constitutional validity of s. 33.1.

[63] With that said, I will now describe with specificity the ways I would find s. 33.1 to be in prima facie infringement of the Charter.

[64] Section 33.1 provides expressly that "[i]t is not a defence to [a violence-based offence] that the accused, by reason of self-induced intoxication, lacked the general intent or the voluntariness required to commit the offence" (emphasis added). The principles of fundamental justice require that voluntariness is an element of every criminal offence. It is therefore contrary to the principle of fundamental justice affirmed in Daviault, at p. 91, to remove the voluntariness element from an offence. It is also contrary to s. 11(d) to convict someone where there is a reasonable doubt about voluntariness.

[65] The Crown does not dispute the importance of voluntariness. It argues instead that the voluntariness inherent in voluntary intoxication supplies the required voluntariness element for the violence-based charges. With respect, the Crown's reliance on the voluntariness of intoxication is misplaced. The purpose of the principle of voluntariness is to ensure that individuals are convicted only of conduct they choose. What must be voluntary is the conduct that constitutes the criminal offence charged, in this case, the assaultive acts by Mr. Chan. Without those assaultive [page376] acts, his voluntary intoxication would be benign. The converse is not so. It is an offence to engage in assaultive acts, even without voluntary intoxication. Clearly, the prohibited conduct that constitutes the offences Mr. Chan is charged with are the assaults, not the self-induced intoxication, and it is the assaults to which voluntariness must attach to satisfy the Charter.

[66] Case law is clear on this point. The Supreme Court of Canada has consistently affirmed that voluntariness must be linked to the prohibited conduct. As LeBel J. put it in Bouchard-Lebrun, at para. 45, it is unfair to convict "an accused who did not voluntarily commit an act that constitutes a criminal offence" (emphasis added). In R. v. Théroux, 1993 CanLII 134 (SCC), [1993] 2 S.C.R. 5, [1993] S.C.J. No. 42, at p. 17 S.C.R., McLachlin J. (as she then was), in speaking of the elements of the crime, said "the act must be the voluntary act of the accused for the actus reus to exist". In his dissenting reasons from Rabey, at p. 522 S.C.R., Dickson J. (as he then was) spoke of the "basic principle that the absence of volition in respect of the act involved is always a defence to a crime" (emphasis added). This passage was subsequently quoted by LaForest J., writing for the majority, in R. v. Parks, 1992 CanLII 78 (SCC), [1992] 2 S.C.R. 871, [1992] S.C.J. No. 71, at p. 896 S.C.R. The act involved in a violence-based offence is the act of violence. The principle of voluntariness is not satisfied by relying on the voluntariness of conduct other than the act that constitutes the criminal offence charged.

[67] The decision in R. v. Penno, 1990 CanLII 88 (SCC), [1990] 2 S.C.R. 865, [1990] S.C.J. No. 96, relied upon by the Crown, does not establish otherwise. Penno dealt with a constitutional challenge to the offence of care or control of a motor vehicle while impaired. The constitutional challenge in that case was untenable because the accused argued that significant impairment should be a defence to the charge, even though impairment is an element of the offence. The court divided in explaining why that constitutional challenge had to fail. However, most of the judges found that since impairment is not only an element of the offence, but also the gravamen of the offence, the voluntariness principle is satisfied by requiring voluntary impairment. The current constitutional challenge differs. The gravamen of the offences Mr. Chan is charged with is not impairment, but his assaultive behaviour, and he is not attempting to convert an element of the offences charged into a defence. The inapplicability of Penno is underscored by the fact that in Daviault, at p. 102 S.C.R., Cory J. cited Penno but nonetheless decided that the Leary rules would contravene the principle of voluntariness.

[68] Moreover, Wilson J., the only judge to address the point in Penno, said that the reasoning in Penno does not apply for offences where intoxication is not made part of the actus reus but [page377] is relevant only to assess the presence of mens rea: at pp. 891-92 S.C.R. When speaking of offences where intoxication is not an element of the offence, she reaffirmed her position in Bernard that the defence of non-mental disorder automatism will be a defence: Penno, at pp. 889-90 S.C.R.

[69] I do not accept the Crown's attempt to overcome the problem that the principle of voluntariness presents by arguing that s. 33.1 creates a new and different mode of committing all Criminal Code offences that "include as an element an assault or any other interference or threat of interference by a person with the bodily integrity of another person'". In effect, the Crown's position is that s. 33.1 adds new, alternative elements to those offences, which permit conviction based on voluntary intoxication, even in the absence of the mens rea specified in the affected sections. On this basis, the voluntariness of the intoxication satisfies the voluntariness requirement.

[70] In support of this interpretation, the Crown relies on the language of s. 33.1(3), which provides:

[71] I cannot agree. Read in context and in its entirety, s. 33.1 does not create a new mode of committing violent offences. The opening words of s. 33.1(3), "This section", are a reference to s. 33.1 as a whole, which begins by stating its function in s. 33.1(1): "It is not a defence to an offence referred to in subsection (3)" (emphasis added). The wording of s. 33.1 removes a defence. This is as expected, since Parliament enacted s. 33.1 as a direct response to a common law rule that recognized involuntariness as a defence.

[72] Moreover, if the function of s. 33.1 was to amend the elements of those offences, one would expect it to be in Part VIII of the Criminal Code, "Offences Against the Person and Reputation", where those offences are found. Instead, the section is placed alongside the defences addressed in the Criminal Code.

[73] Quite plainly, Parliament did not pass s. 33.1 as a "one fell swoop" amendment to a raft of offences. It was passed to eliminate the defence of non-mental disorder automatism for the offences referenced.

[74] Even if s. 33.1 could somehow be interpreted as creating a parallel cast of offences, this would not solve the voluntariness problem. The act prosecuted would remain the commission of the prohibited act specified in the offence charged. As explained, in this [page378] case, the prohibited act is Mr. Chan's assaultive behaviour, as the act of voluntary intoxication is benign without such behaviour.

[75] The trial judge was correct to find that s. 33.1 contravenes ss. 7 and 11(d) of the Charter because it bypasses the requirement of voluntariness, which is a principle of fundamental justice.

[76] In Morrison, at para. 51, Moldaver J., for the majority of the Supreme Court of Canada, recently reaffirmed the s. 11(d) principle relied upon in Daviault:

[77] As Daviault recognizes, at p. 91, substituting voluntary intoxication for the required elements of a charged offence violates s. 11(d) because doing so permits conviction where a reasonable doubt remains about the substituted elements of the charged offence. As the trial judge pointed out in this case, that is the unconstitutional effect of s. 33.1 on Mr. Chan. It purports to permit Mr. Chan to be convicted of manslaughter and aggravated assault without proof of the mental state required by those offences, namely, the intention to commit the assaults.

[78] Of course, if everyone who becomes voluntarily intoxicated necessarily has the intention to commit the charged offences, this constitutional problem would not arise. By proving Mr. Chan's voluntary intoxication, the Crown would inexorably or inevitably also be proving his intention to commit the assaults that supported his manslaughter and aggravated assault convictions. Permitting the Crown to rely on voluntary intoxication in these circumstances would not leave a reasonable doubt about the required elements of the charged offences: R. v. Vaillancourt, 1987 CanLII 2 (SCC), [1987] 2 S.C.R. 636, [1987] S.C.J. No. 83, at p. 656 S.C.R.; R. v. Whyte, 1988 CanLII 47 (SCC), [1988] 2 S.C.R. 3, [1988] S.C.J. No. 63, at pp. 18-19 S.C.R.; Daviault, at pp. 90-91; and Morrison, at paras. 52-53. This argument is not available to the Crown, since proving voluntary intoxication does not necessarily or even ordinarily prove the intention to commit assaults, let alone the assaults charged. The materials [page379] before us from the Standing Committee that was considering Bill C-72 emphasize the correlation between intoxication (particularly alcohol intoxication) and violence, and that link cannot be questioned. However, that link falls far short of showing that those who become intoxicated intend to commit assaults. By enabling the Crown to prove involuntary intoxication instead of intention to assault, s. 33.1 relieves the Crown of its burden of establishing all the elements of the crimes for which Mr. Chan was prosecuted, contrary to s. 11(d) of the Charter.

[79] Section 33.1 also infringes s. 7 of the Charter by enabling the conviction of accused persons who do not have the constitutionally required level of fault for the commission of a criminal offence. The Crown argues that the fault inherent in voluntary intoxication suffices where a person commits an act "that includes as an element an assault or any other interference or threat of interference by a person with the bodily integrity of another person". I do not agree.

[80] In R. v. Creighton, 1993 CanLII 61 (SCC), [1993] 3 S.C.R. 3, [1993] S.C.J. No. 91, at pp. 61-62 S.C.R., the Supreme Court of Canada held that where an offence provides no other mens rea or "fault" requirement, the Crown must at least establish "penal negligence" to satisfy the principles of fundamental justice. Put otherwise, penal negligence is the minimum, constitutionally-compliant level of fault for criminal offences. The general intent offences Mr. Chan was charged with have never been found to require more than the minimum level of fault. Nor is there any reason to conclude that they fall within the "small group of offences" that require a purely subjective standard of fault: Morrison, at para. 75. The standard of penal negligence is therefore the appropriate measure for testing the constitutional validity of s. 33.1, which modifies the fault standard for violence-based offences committed while voluntarily intoxicated.

[81] Indeed, s. 33.1 is built on a theory of negligence. As the Preamble confirms, and the Crown arguments before us suggest, the underlying theory of fault supporting s. 33.1 rests in the irresponsibility of self-induced intoxication and the "close association between violence and intoxication": see Preamble to Bill C-72. Section 33.1 also draws on the language of negligence, referring to a marked departure from reasonable standards of care.

[82] The instant question, then, is whether the fault imposed by s. 33.1 satisfies the penal negligence standard? It does not. [page380]

[83] In Creighton, at p. 59 S.C.R., the Supreme Court of Canada defines penal negligence as negligence that constitutes a marked departure from the standard of a reasonable person. The concept of negligence that girds this standard, which is common to the tort of negligence, operates as an objective measure that involves an assessment of the relationship between an act or omission and a damaging consequence: Mustapha v. Culligan Canada Ltd., [2008] 2 S.C.R. 114, [2008] S.C.J. No. 27, 2008 SCC 27, at paras. 6-15. "Negligence" is not based on whether the person intended or foresaw the damaging consequence, but on whether a reasonable person would have foreseen and avoided the risk that the damaging consequence could occur by not engaging in the allegedly negligent act or omission. If this is so, civil negligence is established. For penal negligence to exist so that criminal liability can be imposed, the relevant risk must be reasonably foreseeable such that it not only falls below standards of ordinary prudence to engage in the risky behaviour but doing so amounts to a marked departure from standards of ordinary prudence: Creighton, at p. 59 S.C.R. Section 33.1 fails to meet this standard in several ways.

[84] First, s. 33.1 does not require a foreseeability link between voluntary intoxication and the relevant consequence, the act of violence charged. In Bouchard-Lebrun, at para. 89, LeBel J. set out the elements of s. 33.1:

[85] Note that on this authoritative description of the elements of s. 33.1, there is no prescribed link between the voluntary intoxication and the violent act. It does not matter how unintentional, non-wilful, unknowing, or unforeseeable the interference with bodily integrity or threatening is. So long as these components each occur, s. 33.1 operates. This is problematic because without a foreseeable risk arising from the allegedly negligent act, negligence cannot be established, and without negligence, the minimum constitutional standard of penal negligence cannot be met.

[86] Second, even if s. 33.1 had required such a link, the charged violent behaviour is not invariably going to be a foreseeable risk of voluntary intoxication, yet s. 33.1 will nonetheless enable conviction. Cory J. made this point in the context of the sexual assault charge before him, in Daviault, at p. 91 S.C.R.:

[87] Mr. Chan's case illustrates the point. A reasonable person in Mr. Chan's position could not have foreseen that his self-induced intoxication might lead to assaultive behaviour, let alone a knife attack on his father and his step-mother, people he loved.

[88] Third, the normative element of penal negligence -- that the allegedly negligent conduct be a marked departure from the standards of a reasonable person -- is absent. It is important to appreciate that the voluntary intoxication required by s. 33.1 does not require an accused person to intend to become intoxicated to the point of automatism, or even to become extremely intoxicated. It is enough to meet the elements of s. 33.1 that a person takes a substance intending to become intoxicated: R. v. Vickberg, 1998 CanLII 15068 (BC SC), [1998] B.C.J. No. 1034, 16 C.R. (5th) 164 (S.C.), at para. 68. This is made clear in Bouchard-Lebrun where LeBel J., recognized, at para. 91, that the self-induced intoxication requirement of s. 33.1 is met even where a voluntary choice to become intoxicated produces abnormal effects. The implication is that a decision to become intoxicated to any degree is enough to trigger s. 33.1, even where the accused person cannot reasonably expect that, as a result of that intoxication, they may become unaware of their behaviour or incapable of consciously controlling their behaviour.

[89] Indeed, the Crown before us goes further. It contends that a person who takes a prescription drug for health-related reasons, and who knows or should know that the drug carries the risk of intoxicating side effects, is under self-induced intoxication if intoxication happens to occur. Relying on non-s. 33.1 cases, the Crown also contends that a person will be voluntarily intoxicated, as in Mr. Sullivan's case, if they take an intoxicating substance not to become intoxicated, but in an attempted suicide: R. v. Turcotte, [2013] J.Q. no 10269, 2013 QCCA 1916, [2013] R.J.Q. 1743, leave to appeal refused [2014] S.C.C.A. No. 7; R. v. Honish, [1991] A.J. No. 1057, 1991 ABCA 304, 120 A.R. 223 (C.A.), at para. 9, aff'd 1993 CanLII 156 (SCC), [1993] 1 S.C.R. 458, [1993] S.C.J. No. 12.

[90] I will leave aside whether the reach of s. 33.1 goes as far as the Crown suggests and focus exclusively on those who intend to become intoxicated, including those who intend for their intoxication to be no more than mild. The notion that it is a marked departure from the standards of the norm to become intoxicated, let alone mildly intoxicated, is untethered from social reality, particularly in a nation where the personal use of cannabis has just been legalized. Voluntary mild intoxication is not uncommon. Whatever [page382] one may think of voluntary mild intoxication, it is difficult to accept that it is a marked departure from the norm.

[91] Finally, even if moral fault can be drawn from voluntary intoxication, it is far from self-evident as a normative proposition that such intoxication is irresponsible enough to substitute for the manifestly more culpable mental states provided for in the general intent offences, such as intention or recklessness relating to sexual assault.

[92] It appears from s. 33.1(2) that Parliament attempted to overcome these challenges by using the language of marked departure and by referencing the standards of reasonable persons. Subsection 33.1(2) provides:

[93] I do not accept the submission made by the intervener, LEAF, that s. 33.1 satisfies minimum standards of constitutional fault because it describes an adequate standard of fault. Whether minimum standards of constitutional fault are met depends on the reach of the section, not the language Parliament uses to describe the level of fault it seeks to impose. For the reasons described, the reach of s. 33.1 does not comply with minimum standards of constitutional fault.

[94] This problem is not overcome by conceiving of the violent act itself as the marked departure, as expressed in the elements of s. 33.1, which were laid out by LeBel J. in Bouchard-Lebrun, at para. 89. This is because moral fault cannot come from a consequence alone. Instead, in the case of negligence, "the mental fault lies in failure to direct the mind to a risk which the reasonable person would have appreciated": Creighton, at p. 58 S.C.R. If a consequence that society judges to be a marked departure from the norm could ground criminal liability, the law would countenance criminal fault based on absolute liability, which would itself violate the Charter: see Reference re Motor Vehicle Act (British Columbia) Section 94(2), 1985 CanLII 81 (SCC), [1985] 2 S.C.R. 486, [1985] S.C.J. No. 73.

[95] The trial judge was therefore correct in finding that s. 33.1 violates the Charter in three distinct ways: (a) a voluntariness breach of ss. 7 and 11(d); (b) an improper substitution breach of s. 11(d); and (c) a mens rea breach of s. 7. Since s. 33.1 is in prima [page383] facie violation of ss. 7 and 11(d) of the Charter in these ways, it is of no force or effect unless the Crown can demonstrate, pursuant to s. 1 of the Charter that s. 33.1 is a "reasonable limit" "prescribed by law as can be demonstrably justified in a free and democratic society". The trial judge found that the Crown met this burden, and therefore, dismissed Mr. Chan's Charter challenge to s. 33.1.

[96] With respect, I would find that the trial judge committed several errors in coming to this conclusion. Most significantly, the trial judge misstated the object of s. 33.1. The purposes he ascribed to s. 33.1 were too broad. These errors are critical because, as I will explain, the trial judge's mistaken determination of purpose tainted each stage of his s. 1 analysis, contributing to errors in his rational connection, minimal impairment, and overall proportionality analysis.

[97] I would also find that s. 33.1 cannot be justified under s. 1. Section 1 analysis is grounded in a contextual application of the framework set out in R. v. Oakes, 1986 CanLII 46 (SCC), [1986] 1 S.C.R. 103, [1986] S.C.J. No. 7. In R.J.R.-MacDonald Inc. v. Canada (Attorney General), 1995 CanLII 64 (SCC), [1995] 3 S.C.R. 199, [1995] S.C.J. No. 68, at paras. 126-130, the Supreme Court of Canada refined without altering the framework for establishing a reasonable limitation finding under s. 1. The Crown must demonstrate:

[98] Section 33.1 would be of no force or effect if the Crown has failed to demonstrate any of these components on a balance of probabilities. I would find that the Crown has not demonstrated the rational connection, minimal impairment, or the proportionality required to save the provision.

[99] In coming to this conclusion, I recognize that courts are to approach constitutional challenges, including s. 1 evaluations, with a "posture of respect" to Parliament: Mills, at para. 56. I [page384] also recognize Parliament's core competency in creating criminal offences. However, courts have core competency in identifying constitutional principles that determine the proper reach of criminal liability in our free and democratic society, and the responsibility to protect those principles from unconstitutional laws: Reference re Motor Vehicle Act (British Columbia) Section 94(2), at para. 15. As Vertes J. observed in R. v. Brenton, "deference is not the same thing as merely taking Parliament's choice at face value. That would be an abdication of [judicial] responsibility": 1999 CanLII 4334 (NWT SC), [1999] N.W.T.J. No. 113, 180 D.L.R. (4th) 314 (S.C.), at para. 78, rev'd for other reasons [2001] N.W.T.J. No. 14, 2001 NWTCA 1, 199 D.L.R. (4th) 119. Even after due deference is accounted for, Parliament's choice in enacting s. 33.1 cannot be demonstrably justified in a free and democratic society.

[100] The Crown argued before the trial judge, and on appeal, that s. 33.1 has two pressing and substantial purposes: (1) "holding individuals accountable for intoxicated violence"; and (2) "protecting the security of the person and equality rights of others, particularly women and children, from violent crimes at the hands of intoxicated offenders". The trial judge accepted that these stated purposes accurately reflect the object of s. 33.1 and that both are pressing and substantial purposes, satisfying the first Oakes requirement.

[101] I agree that Parliament did have an "accountability purpose" and a "protective purpose" in mind. However, the Crown expresses these purposes too generally, and the trial judge erred in following the Crown's lead. The accountability purpose and the protective purpose are more specific than the Crown and the trial judge conceive. Stated properly, the accountability purpose is to hold individuals who are in a state of automatism due to self-induced intoxication accountable for their violent acts. The protective purpose is to protect potential victims, including women and children, from violent acts committed by those who are in a state of automatism.

[102] I accept that the purposes as stated by the Crown find support in the Preamble to s. 33.1. I also recognize that Parliament is entitled to identify its legislative objectives in a statutory preamble, and that those stated objectives must be considered by courts undertaking s. 1 analysis. However, there are constitutional principles that courts must respect in identifying the object of legislation under a [page385] s. 1 analysis. Parliamentary declarations of purpose must be measured against those principles so that the task of identifying whether the object of legislation is constitutionally sound is not delegated to Parliament. As Wagner C.J.C. stated in Frank v. Canada (Attorney General), [2019] 1 S.C.R. 3, [2019] S.C.J. No. 1, 2019 SCC 1, at para. 46, "the integrity of the justification analysis requires that the legislative objective be properly stated": see also Bedford, at para. 78; Carter, at para. 77. When those principles are applied here, it is apparent that the Crown's stated purposes cannot be accepted, and the purposes stated in the Preamble must be refined.

[103] First, the Supreme Court of Canada has repeatedly stressed "[t]he critical importance of articulating the measure's purpose at an appropriate level of generality": Frank, at para. 46. This is because "[t]he relevant objective is that of the infringing measure": Frank, at para. 46. As McLachlin C.J.C. explained in R.J.R.-MacDonald Inc., at para. 144, this must be so "since it is the infringing measure and nothing else which is sought to be justified".

[104] Put otherwise, since the Crown is obliged to demonstrate the need for the infringement under s. 1, the purpose it relies upon should relate to that infringement. Here, the infringing measure, s. 33.1, does not address the prosecution of intoxicated offenders generally. It applies only to those who commit violence-based offences while in a state of automatism due to self-induced intoxication. Properly stated, the object of s. 33.1 must be related to these offenders, and not to intoxicated violent offenders generally.

[105] In Sauvé v. Canada (Chief Electoral Officer), [2002] 3 S.C.R. 519, [2002] S.C.J. No. 66, 2002 SCC 68, McLachlin C.J.C. expanded on this point. She said that "[t]o establish justification, one needs to know what problem the government is targeting, and why it is so pressing and important that it warrants limiting a Charter right": Sauvé, at para. 24. In Frank, at paras. 129-30, Côté and Brown JJ., dissenting, but not on this point, counseled courts to look at the state of the law prior to the impugned legislation, and the scope that the legislature sought to regulate with the impugned law. It cannot be said that the government was targeting the general problem of intoxicated violence when it passed s. 33.1. When s. 33.1 was passed, the general problem of intoxicated violence had already been targeted by the Leary rules, as modified in Daviault, which s. 33.1 leaves untouched. Instead, the scope of s. 33.1 makes clear that it targets the one exception to the Leary rules created in Daviault, namely, violent offences committed by those who are in a state of automatism due to self-induced intoxication. It is an overstatement to claim that the mission of s. 33.1 is directed at intoxicated violence generally. [page386]

[106] It is important to avoid overstating legislative objectives, as the Crown and trial judge have done. McLachlin C.J.C. cautioned in R.J.R.-MacDonald Inc., at para. 144, that if the objective is stated too broadly, its importance may be exaggerated, and the entire s. 1 analysis compromised. As she pointed out in Alberta v. Hutterian Brethren of Wilson Colony, [2009] 2 S.C.R. 567, [2002] S.C.J. No. 66, 2009 SCC 37, at para. 76, "the first three stages of Oakes are anchored in an assessment of the law's purpose".

[107] The issue now under consideration demonstrates the importance of stating the purpose accurately. As Spies J. noted in McCaw, at para. 31, in all four of the cases to save s. 33.1 under s. 1, the courts accepted that the objective of s. 33.1 is consistent with the Preamble.² None of the six cases that struck down s. 33.1 did so.³ They each recognized that s. 33.1 is not targeted at alcohol-induced violence in general, but at the uncommon circumstance of violence committed by offenders while in a state of automatism as the result of self-induced intoxication.

[108] Just as it is perilous to overstate the objective of challenged legislation, it is perilous to understate that objective when approaching s. 1. It understates the objective of s. 33.1 to accept, as some courts have, that the real purpose of s. 33.1 is to remove the narrow defence in Daviault: see Dunn, at para. 34; Brenton, at paras. 102-103; and McCaw, at para. 129. Casting the object of s. 33.1 in this way masks the underlying reason why Parliament wanted to remove that narrow defence, and it improperly confuses the means of the legislation with its purpose, which the Supreme Court of Canada has held to be erroneous in R. v. J. (K.R.), [2016] 1 S.C.R. 906, [2016] S.C.J. No. 31, 2016 SCC 31, at para. 63.

[109] When McLachlin C.J.C. and Major J., said, in Harper v. Canada (Attorney General), [2004] 1 S.C.R. 827, [2004] S.C.J. No. 28, 2004 SCC 33, at para. 25 (dissenting on other grounds) that, "the proper question at this stage of the analysis is whether the Attorney General has asserted a pressing and substantial objective" (emphasis in original), they were not saying that the principles I have just identified should be forgotten and the s. 1 analysis is to be based solely on the government's articulation of the objective. [page387] They were making the point that in judging whether a purpose is pressing and substantial, evidence is not required, and courts may consider the identified objective using common sense alone to determine if it is pressing and substantial.

[110] Accordingly, as the principles I have identified verify, in conducting a s. 1 analysis, courts must look at the substance of what is being done to determine the purpose of the legislation. On this basis, this court refined the Attorney General's asserted purpose in Longley v. Canada (Attorney General) (2007), 88 O.R. (3d) 408, 2007] O.J. No. 4758, 2007 ONCA 852, at para. 49, and the Supreme Court of Canada recently did so in Frank, at paras. 49 and 54. This enterprise is not about passing judgment on whether the Crown acted in good faith in describing the purpose as it did. It is about ensuring that the constitutional issues raised are addressed in context.

[111] Properly stated, the underlying purposes or objectives of s. 33.1 are: (1) to hold individuals who are in a state of automatism due to self-induced intoxication accountable for their violent acts [the "accountability purpose"]; and (2) to protect potential victims, including women and children, from violence-based offences committed by those who are in a state of automatism due to self-induced intoxication [the "protective purpose"].

[112] The accountability purpose is an improper "purpose" for s. 1 evaluation. Therefore, it cannot serve as a pressing and substantial purpose.

[113] The reason can be stated simply. The constitutional principles at issue define when criminal accountability is constitutionally permissible, given entrenched, core values. To override principles that deny accountability, for the purpose of imposing accountability, is not a competing reason for infringing core constitutional values. It is instead a rejection of those values. It cannot be that a preference for other values over constitutionally entrenched values is a pressing and substantial reason for denying constitutional rights. The point can be put more technically by examining two principles that govern s. 1 evaluation.

[114] First, legislation is unconstitutional if its purpose is unconstitutional: R. v. Big M. Drug Mart Ltd., 1985 CanLII 69 (SCC), [1985] 1 S.C.R. 295, [1985] S.C.J. No. 17, at p. 333 S.C.R. Since the Charter principles at stake here describe when it is unconstitutional to hold someone criminally accountable (i.e., in the absence of voluntariness or [page388] penal negligence), passing legislation to impose criminal accountability despite those principles is an unconstitutional purpose. A purpose cannot at once be unconstitutional and a pressing and substantial reason for overriding constitutional rights.

[115] Second, all criminal legislation exists to hold offenders accountable. If accountability could serve as a pressing and substantial objective in criminal cases, the pressing and substantial purpose standard would be met whenever Parliament chooses to criminalize conduct. The Supreme Court of Canada has cautioned against accepting "purposes" that would inoculate any criminal legislation: R. v. Zundel, 1992 CanLII 75 (SCC), [1992] 2 S.C.R. 731, [1992] S.C.J. No. 70, at p. 761 S.C.R.; Sauvé, at para. 24.

[116] For these reasons, the trial judge erred in relying on accountability as a pressing and substantial purpose and in using that purpose to frame the balance of his analysis.

[117] In Daviault, Cory J. concluded that the protective purpose is not a pressing and substantial basis for infringing Charter principles. Given the infrequency of non-mental disorder automatism, there is no pressing need to remove the defence. At pp. 92-93 S.C.R., he explained:

[118] However, this analysis from Daviault is not binding because it addressed the state of the common law, not the constitutionality of s. 33.1.⁴ The "pressing and substantial purpose" holding is, therefore, open for reconsideration, and I am persuaded by my colleague that the existence of a pressing and substantial purpose should not turn solely on the infrequency of the problem addressed. As the tragic outcome in the cases now before this court [page389] demonstrate, even though acts of violence may only rarely be committed by individuals in a state of intoxicated automatism, the consequences can be devastating. This is enough to satisfy me that seeking to protect potential victims, including women and children, from violence-based offences committed by those who are in a state of automatism due to self-induced intoxication is a pressing and substantial purpose.

[119] The rational connection requirement describes the link between the legislative objective and the legislative means chosen to achieve that objective. This rational connection need not be proven on a rigorous scientific basis. A causal connection based on reason or logic may suffice: R.J.R.-MacDonald Inc., at paras. 137, 156. The Crown must establish a reasoned basis for concluding that "the legitimate and important goals of the legislature are logically furthered by the means the government has chosen to adopt": Lavigne v. O.P.S.E.U., 1991 CanLII 68 (SCC), [1991] 2 S.C.R. 211, [1991] S.C.J. No. 52, at p. 291 S.C.R.

[120] The trial judge in Chan held that only the accountability purpose identified by the Crown satisfied this standard. He concluded that the protective purpose, even when expressed as broadly as it was by the Crown, did not.

[121] I agree with the trial judge on the latter point. As the Crown recognized, deterrence is the means s. 33.1 relies upon to achieve its protective purpose. The trial judge was unpersuaded, "as a matter of common sense, that many individuals are deterred from drinking, in the off chance that they render themselves automatons and hurt someone". I share that position. Effective deterrence requires foresight of the risk of the penal consequence. I am not persuaded that a reasonable person would anticipate the risk that, by becoming voluntarily intoxicated, they could lapse into a state of automatism and unwilfully commit a violent act. Even if this remote risk could be foreseen, the law already provides that reduced inhibitions and clouded judgment, common companions of intoxication, are no excuse if a violent act is committed. It is unlikely that if this message does not deter, removing the non-insane automatism defence will do so. Even bearing in mind the admonition in R. v. Malmo-Levine; R. v. Caine, [2003] 3 S.C.R. 571, [2003] S.C.J. No. 79, 2003 SCC 74, at para. 177, to exercise caution in accepting arguments about the ineffectiveness of legislative measures, I am not persuaded that s. 33.1 furthers the public protection purpose. [page390]

[122] Nor can a rational connection be built upon the accountability purpose. I accept that the legislation is effective at achieving accountability, however, for the reasons already explained, accountability cannot be relied upon as a proper objective for s. 33.1. To use the language from Lavigne, a rational connection must be built upon "legitimate and important goals": at p. 291 S.C.R., per Wilson J. Accountability is not a legitimate goal to employ to override Charter rights, which are designed to limit accountability. The trial judge erred in building a rational connection on the accountability objective.

[123] In Morrison, at para. 68, Moldaver J. reaffirmed that "[t]o show minimal impairment, the party seeking to justify the infringement must demonstrate that the impugned measure impairs the right in question 'as little as reasonably possible in order to achieve the legislative objective'". This does not require Parliament to adopt the least restrictive means possible. The issue is whether Parliament could reasonably have chosen an alternative means which would have achieved the identified objective as effectively: R. v. Chaulk, 1990 CanLII 34 (SCC), [1990] 3 S.C.R. 1303, [1990] S.C.J. No. 139, at p. 1341 S.C.R.

[124] The trial judge found s. 33.1 to be minimally impairing. He accepted the Crown's submissions that: (1) s. 33.1 is narrowly tailored because s. 33.1 is confined to violence-based, general intent offences involving self-induced intoxication; (2) Parliament had valid reasons for rejecting the only alternative that would directly achieve the objective of the legislation in a less impairing way; and (3) he should defer to the choice of Parliament.

[125] I have concluded that the trial judge erred in making each of these decisions.

[126] The purported narrow tailoring of s. 33.1 does not provide a basis for a minimal impairment finding, as the identified limitations are not substantial.

[127] By its terms, s. 33.1 is not confined to general intent offences. Section 33.1 prevents self-induced intoxication from being relied upon to establish that the accused "lacked the general intent or the voluntariness required to commit the [violence-based] offence" (emphasis added). On the face of s. 33.1, self-induced intoxicated automatism cannot be used to rebut voluntariness for any violence-based offence, regardless of whether it involves general or specific intent. [page391]

[128] Moreover, s. 33.1 was confined to violence-based offences not to confine its reach but because, as the Preamble and the history of the provision confirms, this is the problem that Parliament was addressing. The mischief Parliament set out to address is covered completely. There is therefore no realistic foundation for the suggestion that the reach of s. 33.1 has been curtailed to achieve restraint.

[129] Finally, and as already explained, the conception of the kind of self-induced intoxication that will undermine an automatism defence is aggressive in its scope. It is not confined to those who choose to become extremely intoxicated and to thereby court the remote risk of automatism. The Crown's position is that anyone consuming an intoxicant, including prescription medication that they know can have an intoxicating effect, is caught, as are those who become intoxicated in the course of suicide attempts.

[130] I would also note that, for those who are caught by s. 33.1, the relevant Charter rights are not merely infringed or compromised. They are denied entirely. I do not agree that s. 33.1 is narrowly tailored.

[131] However, narrow tailoring is not the central concern. Ultimately, minimal impairment is tested not by whether efforts were made to confine its reach, but by whether, given the context, Parliament could reasonably have chosen less intrusive alternative means, which would have achieved the identified objective as effectively. In my view, the Crown has failed to demonstrate that there are not less intrusive reasonable alternatives.

[132] First, I agree with the trial judge that the option of a stand-alone offence of criminal intoxication would achieve the objective of s. 33.1. Making it a crime to commit a prohibited act while drunk is the response Cory J. invited in Daviault, at p. 100 S.C.R., and that was recommended by the Law Reform Commission of Canada: see Recodifying Criminal Law, Report 30, vol. 1 (1986), at pp. 27-28. It is difficult to reject this option as a reasonable alternative given the impressive endorsements it has received.

[133] But would this new offence be equally effective as s. 33.1? Creating such an offence would arguably be more effective in achieving the Preamble objective of protecting against acts of intoxicated violence, as it would serve to deter voluntary intoxication directly and more broadly than s. 33.1 does. It would do so by making the act of intoxication itself the gravamen of the offence, and its reach would not be confined to those who are in a state of automatism because of self-induced intoxication. Instead, its reach would [page392] depend on whether the intoxication was dangerous, as demonstrated by the commission of a violence-based offence.

[134] Certainly, this option would also be less impairing than s. 33.1 since it does not infringe, let alone deny, the Charter rights that s. 33.1 disregards. It would criminalize the very act from which the Crown purports to derive the relevant moral fault, namely, the decision to become intoxicated in those cases where that intoxication proves, by the subsequent conduct of the accused, to have been dangerous.

[135] I do not agree with the trial judge, or the Crown, that Parliament had valid reasons for choosing s. 33.1 instead of this option. Two of the reasons relied upon for doing so are legally invalid and it was an error for the trial judge to accept them. More specifically, the objections that such an offence would: (1) appear to create a sentencing discount for intoxicated offenders; or (2) undermine the object of accountability by suggesting that the accused is not guilty of the violence-based act, are accountability concerns. As indicated, the desire to impose accountability cannot support a reasonable limit on Charter rights that exist to restrict the reach of accountability, such as the Charter rights denied by s. 33.1. In any event, it would not be the offence of intoxicated violence that suggests that the accused is not guilty of the violence-based act. It is the presumption of innocence and the principles of fundamental justice that produce this result.

[136] Nor can the rejection of the criminal intoxication option be justified on the basis that such an offence may have other constitutional problems of its own. I understand that the unconstitutionality of an option would make that option unreasonable, but I cannot accept that a constitutional infringement can be justified as a stratagem for avoiding another possible constitutional infringement.

[137] The alternative option that the Crown has not disproved is to simply permit the Daviault decision to operate. By design, the non-mental disorder automatism defence is difficult to access. As with other defences, if there is no air of reality to the defence based on the evidence, it should not be considered: Stone, at paras. 166-68. It is also a reverse onus defence, and it requires expert evidence: Daviault, at p. 101 S.C.R. If the defence is not established on the balance of probabilities, it fails: Stone, at para. 179. Indeed, it may well have failed for Mr. Daviault had the complainant not died before his retrial. According to evidence that Parliament has accepted, alcohol intoxication is not capable, on its own, of inducing a state of automatism: see Preamble of Bill C-72. Had similar evidence been presented and accepted at Mr. Daviault's retrial, he would have been convicted. [page393]

[138] Moreover, even in those few cases where the accused might succeed in demonstrating automatism as the result of the voluntary consumption of intoxicants, the accused may not be acquitted. If the accused is unable to establish that the cause of the automatism was not a disease of the mind, which it will be if the automatism is internally caused or there is a continuing danger of further episodes of automatism, the accused will not be acquitted, but found not criminally responsible on account of mental disorder: Stone, at paras. 197-217. The accused would then be subject to a disposition hearing driven by public safety considerations.

[139] I do not accept the Crown's submission that accepting this "do nothing" option cannot operate as a more minimally impairing strategy because "it directly subverts Parliament's goal by allowing extremely intoxicated violent offenders to escape liability". Again, this is an accountability argument and, as I have indicated, given that the principles of fundamental justice at stake exist to define the constitutional preconditions to criminal accountability, the desire to impose accountability is itself an unconstitutional purpose.

[140] This submission also materially understates the effect of the common law after Daviault. As demonstrated, in the few cases where there will be an air of reality to the concern that extreme intoxication has led to automatism and then to violence, the prospects of escaping liability are slim. I have already expressed my view that it is unrealistic to think that s. 33.1 adds any meaningful deterrence augmentation to the Leary rules, as modified in Daviault. Realistically, who would choose to consume intoxicants because they have reasoned that, if all goes wrong, they will have the non-mental disorder automatism defence? However, assuming for the sake of the exercise that s. 33.1 could have some additional deterrent effect, one would think that the unlikelihood of the common law defence succeeding would have a comparable deterrent effect.

[141] In the circumstances, I am satisfied that the Crown has not disproved that the Daviault regime is not a reasonable and equally effective but less impairing alternative to s. 33.1, in protecting potential victims from violence committed by those who are in a state of automatism as the result of self-induced intoxication.

[142] Third, I disagree with the trial judge's reliance on deference to support his finding that minimal impairment had been demonstrated. The trial judge was correct to turn his mind to this. The context-driven inquiries that s. 1 entails generally call for deference, particularly in examining minimal impairment. However, this is not a case where there is room for the kind of [page394] reasonable disagreement that could trigger deference. The minimal impairment test is simply not met.

[143] I would therefore hold that the trial judge erred in finding that the Crown demonstrated minimal impairment, and find that s. 33.1 is not, in fact, minimally impairing.

[144] Overall proportionality entails the proper identification of the salutary or positive effects of the legislation, and its deleterious or negative effects on the Charter-protected interests at stake. To save legislation that is in prima facie violation of the Charter, the Crown must demonstrate on the balance of probabilities, that there is proportionality between those salutary and deleterious effects.

[145] The trial judge found, in this case, that the Crown had demonstrated overall proportionality. I accept the trial judge's conclusion that "no right is sacrosanct. Each must be considered in context and each may at times bend to other pressing rights or concerns." However, with respect, I would find that he erred in his reasoning, and in the result he achieved.

[146] First, the trial judge's analysis rests heavily on the salutary effects of imposing accountability. As explained, I am persuaded that the accountability purpose cannot be relied upon in the s. 1 evaluation, given that infringing constitutional limits on accountability in order to impose accountability is itself an unconstitutional purpose.

[147] Second, the trial judge predicated his balancing on the generic proposition that "[t]hose who self-intoxicate and cause injury to others are not blameless". He did so without apparent recognition of the expansive grasp of the concept of self-induced intoxication, catching as it does, even those who would fall into a state of automatism after choosing to become mildly intoxicated, and perhaps even those who are complying with a prescribed, medically-indicated drug that they know may cause intoxicating effects. The theory of moral fault that he relied upon cannot be sustained.

[148] Third, the trial judge gave undue weight to the extent to which s. 33.1 provides for the safety of the potential victims, including women and children. As I have indicated, I am persuaded that the protection thesis cannot be supported on a reasoned basis. Viewing the matter realistically, the deterrence that the law achieves must come from the Leary rules, as modified in Daviault, not from the added and remote prospect that if a rare and unforeseen case of automatism should happen to occur and lead to violence, non-mental disorder automatism is off the table. [page395]

[149] Fourth, despite recognizing that the identified Charter infringements are serious, the trial judge minimized their impact by observing that they arise in very few cases. The proper measure is the impact of s. 33.1 on those it affects, not its lack of impact on those it does not affect.

[150] For these reasons, I would conclude that the trial judge erred in applying the overall proportionality test, and I would find that the Crown has failed to demonstrate that overall proportionality is attained.

[151] The deleterious effects of s. 33.1 are profound. Specifically, s. 33.1 enables the conviction of individuals of alleged violence-based offences, even though the Crown cannot prove the requisite elements of those offences, which is contrary to the principles of fundamental justice and the presumption of innocence. It enables the conviction of individuals for acts they do not will. It enables the conviction of individuals of charged offences, even though those individuals do not possess the mens rea required by those offences, or even the minimum level of mens rea required for criminal fault. And it does so, predicated on a theory of moral fault linked to self-induced intoxication, expressed by the Crown before us in language captured in R. v. Decaire, supra, at para. 20: "People who consume alcohol should recognize that continuing to drink after they sense a loss of control of inhibitions, poses a danger to themselves and others." Yet s. 33.1 is not confined to those who set out to become extremely intoxicated. It employs a definition of self-induced intoxication that catches anyone who has consumed an intoxicant, including with restraint or perhaps even for medically-indicated purposes.

[152] Moreover, as Cory J. recognized in Daviault, at p. 87 S.C.R., even leaving aside the other objections I have identified, it is not appropriate to transplant the mental element from the act of consuming intoxicants for the mental element required by the offence charged, particularly where the act of self-inducing intoxication is over before the actus reus of the offence charged occurs. This is what s. 33.1 seeks to do. This transplantation of fault is contrary to the criminal law principle of contemporaneity, which requires the actus reus and mens rea to coincide at some point: see R. v. Williams, [2003] 2 S.C.R. 134, [2003] S.C.J. No. 41, 2003 SCC 41, at para. 35.

[153] Put simply, the deleterious effects of s. 33.1 include the contravention of virtually all the criminal law principles that the law relies upon to protect the morally innocent, including the venerable presumption of innocence.

[154] Only the most compelling salutary effects could possibly be proportional to these deleterious effects. Yet s. 33.1 achieves little. If not entirely illusory, its contribution to deterrence is negligible. I [page396] have already explained that the protective purpose relied upon carries little weight.

[155] The Crown and supporting interveners argue that s. 33.1 has collateral salutary effects, such as: "(i) encouraging victims to report intoxicated violence, (ii) recognizing and promoting the equality, security, and dignity of crime victims, particularly women and children who are disproportionately affected by intoxicated violence, and (iii) avoiding normalizing and/or incentivizing intoxicated violence".

[156] I see no reasoned basis for concluding that victims who would have reported intoxicated violence would be unlikely to do so because of the remote possibility that a non-mental disorder automatism defence could be successfully raised, or that s. 33.1 plays a material role in preventing the normalization and incentivization of intoxicated violence. Section 33.1 addresses a miniscule percentage of intoxicated violence cases.

[157] As for recognizing and promoting the equality, security, and dignity of crime victims, it is obvious that those few victims who may see their offenders acquitted without s. 33.1 will be poorly served. They are victims, whether their attacker willed or intended the attack. However, to convict an attacker of offences for which they do not bear the moral fault required by the Charter to avoid this outcome, is to replace one injustice for another, and at an intolerable cost to the core principles that animate criminal liability.

[158] What, then, of the benefits of imposing accountability on those who are in a state of non-mental disorder automatism when they commit violent acts? If I am mistaken, and this is a proper s. 1 consideration, would the benefit of doing so alter the balance? Not in my view.

[159] My colleague describes the accountability benefit as ensuring that those who are in a state of self-induced intoxicated automatism are subject to the same penal consequences for violent acts as those whose state of intoxication fall just below a state of automatism. With respect, the move to accountability should not be seen as an exercise in eliminating a distinction based on degree. The material distinctions are between: those who act wilfully and those who do not; those who are proved to have the mens rea for the charged offence and those who do not; and those who have the constitutionally minimum level of fault and those who do not. When balancing the competing interests, it must be remembered that the decision to impose accountability is in direct contravention of the relevant Charter principles. Even if accountability is a proper s. 1 consideration, the benefits it brings must be seen in that light, and its value diminished accordingly. [page397] The benefit of accountability is not, alone or when combined with other salutary effects, proportional to its deleterious effects

[160] Recently, in Morrison, at para. 72, the Supreme Court of Canada held that the promise of additional convictions for the serious offence of child luring could not outweigh the deleterious effect of "sweeping in accused persons" whose mens rea "may be the subject of reasonable doubt". The circumstances are distinguishable, but the outcome is the same. With very little true gain, Parliament has attempted to cast aside the bedrock of moral fault. I would find that the Crown has not shown that s. 33.1 achieves overall proportionality.

[161] I would conclude that the Crown has not demonstrated that s. 33.1 is a demonstrably justifiable limit on the Charter rights at stake, in a free and democratic society. Accordingly, I would declare s. 33.1 to be of no force or effect, pursuant to s. 52(1) of the Constitution Act, 1982.

[162] Since Mr. Chan should have been provided with the opportunity to invoke the non-mental disorder automatism defence, I would set aside his convictions and order a new trial.

[163] Mr. Chan urges that the proper outcome is an acquittal. He contends that since the trial judge found Mr. Chan to be incapacitated, other than by reason of mental disorder, the automatism defence is satisfied. I do not agree.

[164] The trial judge made no finding that Mr. Chan was not acting voluntarily. Instead, he found that as a result of psychosis induced by intoxication, Mr. Chan was incapable of knowing that his actions would be considered wrong according to moral standards of reasonable members of society. This is not a finding of non-mental disorder automatism. A person can lack the capacity to know their acts are wrong, yet still voluntarily choose to engage in those acts.

[165] Mr. Chan sought to overcome the distinction I have identified by relying on Cory J.'s references in Daviault to "extreme intoxication akin to automatism or insanity". Mr. Chan argues that non-mental disorder automatism, as described in Daviault, encompasses his situation because his mental state was akin to "insanity" or mental disorder, even if caused by extreme intoxication. I do not accept this submission. That language was not intended to extend the non-mental disorder automatism defence beyond cases of automatism. In Daviault, at p. 100 S.C.R., Cory J. emphasized that: [page398]

[166] Since the trial judge did not consider whether Mr. Chan had reached the stage of automatism, he is entitled to a new trial, not an acquittal.

[167] I would therefore allow Mr. Chan's appeal, set aside his convictions, and order a new trial.

[168] As a result of the COVID-19 emergency, the panel relieved Mr. Chan from the term of his bail that requires him to surrender into custody prior to this decision being released.

[169] Mr. Sullivan's extreme intoxication resulted from his ingestion of the drug, Wellbutrin. The Wellbutrin was prescribed to help him stop smoking. Psychosis is one of its known side effects. From the time Mr. Sullivan began taking and occasionally abusing Wellbutrin, he experienced episodes where he believed aliens he called "Archons" were living in the condominium he shared with his mother.

[170] On December 1, 2013, after ingesting between 30 to 80 of the Wellbutrin tablets in a suicide attempt, he had a profound break with reality. He believed he had captured an Archon in the condominium living room. He brought his mother into the living room to show her. As she tried to assure him that there was nothing in the room, believing her to be an alien, he attacked her, stabbing her several times with two kitchen knives.

[171] During the attack, his mother screamed, "David, I'm your mother". Mr. Sullivan dropped the knives and ran to a bedroom. Emergency services were called. When the police arrived, Mr. Sullivan was outside of the apartment complex screaming incoherently and running erratically. His mother survived the attack but died of unrelated causes before trial.

[172] At his trial, it was not disputed that Mr. Sullivan was acting involuntarily when he stabbed his mother. Mr. Sullivan attempted to rely on the defence of non-mental disorder automatism but did not challenge the constitutional validity of s. 33.1. He argued, instead, that s. 33.1 did not apply in his case since his intoxication was not voluntary. In the alternative, he invoked the mental disorder defence.

[173] The trial judge agreed that Mr. Sullivan's attack against his mother was involuntary. In considering the implications of [page399] that finding, he began with the mental disorder defence, which is presumed to apply where automatism has been established: Stone, at para. 199. Taking a "holistic approach" to determine the nature of the automatism, the trial judge concluded that the cause of Mr. Sullivan's automatism was external, and that he did not pose a continuing danger. The trial judge found that "this is one of the rare cases where automatism was not caused by mental disorder", but by intoxication. He therefore rejected the mental disorder defence under s. 16 of the Criminal Code.

[174] The trial judge then considered the non-mental disorder automatism defence. He found that this defence was prevented by s. 33.1 because Mr. Sullivan's intoxication had been voluntary. The trial judge applied the following test in making that determination: "Voluntary intoxication means that Mr. Sullivan consumed Wellbutrin when he knew or had reasonable grounds to believe that it might cause him to be impaired."

[175] Accordingly, Mr. Sullivan was convicted of aggravated assault, contrary to Criminal Code, s. 268(1), and using a weapon, a knife, in committing an assault, contrary to Criminal Code, s. 267(a).

[176] At his trial, Mr. Sullivan was also convicted of four counts of failing to comply with recognizance orders, relating to post-attack communications he had with his sister, in breach of a December 4, 2013 non-communication order prohibiting him from contacting her. Those calls occurred between December 29, 2013 and January 5, 2014.

[177] Mr. Sullivan attempted to defend these charges by maintaining that he was unaware of the communication order. He testified to that effect.

[178] The trial judge did not believe that testimony because of his low credibility and reliability. He then said:

[179] In Mr. Sullivan's conviction appeal,⁵ he submits that he should be permitted to raise a Charter challenge to s. 33.1 of the [page400] Criminal Code on appeal for the first time, and that s. 33.1 should be found to be unconstitutional. He also argues that in the course of his judgment, the trial judge erred in defining "voluntary intoxication". Finally, Mr. Sullivan contends that the trial judge erred in finding him guilty of breach of recognizance charges without proof that he knew the terms of his recognizance. He requests verdicts of acquittal on all charges.

[180] The appeal issues can therefore be stated as follows:

[181] Since I would resolve the first two grounds of appeal in the affirmative and they resolve the s. 33.1 issues, the third ground of appeal need not be addressed.

[182] The Crown agrees that the trial judge's reasons disclose that s. 33.1 was the sole basis for Mr. Sullivan's convictions of the violence-based offences. The Crown conceded that if this court declares s. 33.1 unconstitutional in the Chan appeal, Mr. Sullivan's violence-based convictions should be set aside, even though he did not raise the constitutional validity of s. 33.1 at his trial. This concession is obviously correct since Mr. Sullivan's case is still in the system and convictions that depend upon a law that is of no force or effect cannot be upheld on appeal.

[183] Given the conclusion in Chan that s. 33.1 is of no force or effect, I would conclude that the trial judge erred in relying on s. 33.1 and allow this ground of appeal. As conceded by the Crown, the trial judge found Mr. Sullivan to have been in a state of non-mental disorder automatism at the time of the attacks that led to his convictions of aggravated assault and assault using a weapon. [page401]

[184] There is controversy nationally about whether a breach of a term of recognizance contrary to s. 145(3) of the Criminal Code requires a subjective or objective mens rea.⁶ Binding authority of this court in R. v. Legere (1995), 1995 CanLII 1551 (ON CA), 22 O.R. (3d) 89, [1995] O.J. No. 152 (C.A.) has applied a subjective mens rea standard. Therefore, the Crown must establish that the accused had actual knowledge of the condition. Mr. Sullivan's appeal proceeded on this basis.

[185] Mr. Sullivan does not take issue with the trial judge's finding that he had the ability to understand what transpired in court. His issue is with the trial judge's finding that he had the requisite subjective knowledge of the conditions to support his convictions. Specifically, Mr. Sullivan challenges the trial judge's "assumption" that the conditions of the non-communication order would have been discussed at some, if not all, of his court appearances.

[186] I need not decide whether the trial judge erred in making this finding because Mr. Sullivan joined in an agreed statement of facts that contained two relevant passages that confirm his subjective knowledge:

[187] During his testimony, Mr. Sullivan did not resile from the position that the order was imposed by a competent court in his presence. Even if the trial judge should not have found on the evidence that the terms would have been discussed during Mr. Sullivan's multiple court appearances, no miscarriage of justice has [page402] occurred. The trial judge's conclusion that the terms of the recognizance had been communicated to Mr. Sullivan is supported by the uncontested facts. I would therefore reject this ground of appeal.

[188] I would therefore allow Mr. Sullivan's appeal from his convictions of aggravated assault, contrary to Criminal Code, s. 268(1), and using a weapon, a knife, in committing an assault, contrary to Criminal Code, s. 267(a). I would set aside those convictions and substitute verdicts of acquittal. I would reject Mr. Sullivan's appeal of his breach of recognizance convictions and affirm those convictions. Since Mr. Sullivan has already served his sentence on the breach of recognizance offences, there is no need to adjust his sentence to reflect the mixed success of his appeal.

[189] I concur in the result reached by my colleague. I agree that: s. 33.1 of the Criminal Code, R.S.C. 1985, c. C-46 limits the Charter rights of the appellants under ss. 7 and 11(d); the Crown has not met its burden under s. 1 of the Charter of demonstrating that the limits s. 33.1 imposes are "reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society;" and consequently, s. 52(1) of the Constitution Act, 1982 deems s. 33.1 to be "of no force or effect" to the extent of any inconsistency with the Charter.

[190] I concur without reservation with my colleague's reasons as expressed in the overview, his analysis concluding that the Chan trial judge was not bound by precedent to accept the unconstitutionality of s. 33.1, and the disposition.

[191] I have reservations in two areas. First, I agree generally with my colleague's analysis of the limits imposed by s. 33.1 on the ss. 7 and 11(d) rights of the appellants but make additional observations that are especially pertinent to the subsequent s. 1 analysis.

[192] Second, I believe that this court is bound by the Supreme Court's decisions in Bedford and Carter to apply the Bedford/Carter framework to the issue of whether s. 33.1 limits s. 7 Charter rights, in addition to the more traditional analysis my colleague undertakes: Canada (Attorney General) v. Bedford, [2013] 3 S.C.R. 1101, 2013 SCC 72; Carter v. Canada (Attorney General), [2015] 1 S.C.R. 331, 2015 SCC 5.

[193] Finally, I disagree with the substance and tone of my colleague's analysis of s. 1 of the Charter and its application to s. 33.1 of the Criminal Code, while I concur in the result.

[194] I address each reservation in turn. [page403]

[195] R. v. Daviault modified the Leary rule: 1994 CanLII 61 (SCC), [1994] 3 S.C.R. 63, [1994] S.C.J. No. 77; R. v. Leary (1977), 1977 CanLII 2 (SCC), [1978] 1 S.C.R. 29, [1977] S.C.J. No. 39. The Daviault majority held that to be Charter-compliant, evidence of extreme self-induced intoxication must be admissible in defence, whether the offence is one of general intent or specific intent. An accused person who can establish that the offence was committed in a state of automatism resulting from self-induced intoxication, on the balance of probabilities and with the assistance of expert evidence, is entitled to be acquitted.

[196] With respect to these appeals, I note that, but for s. 33.1, the defence of intoxication would have been available to Mr. Chan and Mr. Sullivan on the basis of Daviault. The charges at issue in these appeals are almost all general intent offences. Mr. Chan was convicted of manslaughter, contrary to s. 234 of the Criminal Code, and aggravated assault, contrary to s. 268. Mr. Sullivan was charged with aggravated assault, contrary to s. 268; assault with a weapon, contrary to s. 276(a); possession of weapon for dangerous purpose, contrary to s. 88; and failure to comply with recognizance, contrary to s. 145(3) of the Criminal Code. He was convicted of aggravated assault, assault with a weapon, and breach of recognizance. Only possession of weapon for dangerous purpose is a specific intent offence. The rest are general intent offences.

[197] The trial judge in Chan correctly held that s. 33.1 limits the ss. 7 and 11(d) Charter rights at issue in these appeals by substituting the mental element and thereby bypassing the voluntariness and mental element requirements for criminal convictions and the presumption of innocence on the predicate violent offences. Section 33.1 replicates the same defects in the Leary rule that the Supreme Court corrected in Daviault. The trial judge rightly recognized that s. 33.1 "does the very thing that was . . . held unconstitutional in Daviault", albeit in a narrower compass: at para. 72. His analysis in paras. 46, 48, 72, and 79-80 is particularly trenchant, and I would agree with it.

[198] Section 33.1 tries to sidestep Daviault by substituting the mental element associated with penal negligence for the mental element ordinarily required for the predicate violent acts. But, in Daviault, the Supreme Court found that this type of substitution -- replacing the mental element for sexual assault with the mental element required for intoxication, for example -- was a fatal flaw in the Leary rule. Did the design of s. 33.1 overcome the court's concern? I agree with my colleague that it did not. [page404]

[199] The outcome of these appeals turns on whether the limits s. 33.1 imposes on the appellants' ss. 7 and 11(d) Charter rights can be justified under s. 1.

[200] The trial judge in Chan referred to the Crown's suggestion that, following Bedford and Carter, "the court must measure s. 33.1 against the principles of arbitrariness, overbreadth and gross disproportionality": at para. 92. The trial judge declined to do so because the point had not been fully argued before him, and because Mr. Chan chose to rely on the other fundamental principles discussed above.

[201] Bedford and Carter were raised in these appeals, along with this court's decision in R. v. Michaud in which this court was the first in Canada to uphold a limit on a s. 7 right under s. 1 of the Charter: (2015), 127 O.R. (3d) 81, [2015] O.J. No. 4540, 2015 ONCA 585 (C.A.), leave to appeal refused [2015] S.C.C.A. No. 450. Bedford and Carter reframed the relationship between ss. 7 and 1 of the Charter, as was explained in Michaud, at para. 62. Consequently, the trial judge was, and this court is, obliged to consider the Bedford/Carter reframing.

[202] Before addressing the s. 7 analysis under the Bedford/Carter framework, I consider the restated relationship between ss. 7 and 1 of the Charter.

[203] Section 7 of the Charter is meant to assess "the negative effect on the individual against the purpose of the law, not against societal benefit that might flow from the law" (emphasis in original): Bedford, at para. 121. Section 7 focuses on the relationship between the individual claimant and the law, while s. 1 of the Charter focuses on the relationship between the private impact and the public benefit of the law: Bedford, at paras. 124-29. The balancing function -- "whether the negative impact of the law on the rights of individuals is proportional to the pressing and substantial goal of the law in furthering the public interest" -- is addressed only in the s. 1 Charter analysis, after the claimant has established the s. 7 limit: Bedford, at para. 125.

[204] I would therefore not give effect to the submission of the intervener LEAF that s. 7 of the Charter requires "internal balancing" in identifying the relevant principles of fundamental justice for consideration. It is the court's task under s. 1 of the Charter, not under s. 7, to carry out any balancing between the accused's [page405] interests and the public interest that LEAF asserts -- equality and the human dignity of women and children who are disproportionally victimized by intoxicated offenders.

[205] I next carry out the s. 7 analysis under the Bedford/Carter framework before conducting the s. 1 analysis.

[206] The analytical framework established in the s. 7 jurisprudence requires an assessment of three negative "principles of fundamental justice": arbitrariness, overbreadth, and gross disproportionality: Bedford, at paras. 94, 96. In Bedford, the Supreme Court described these principles as "failures of instrumental rationality": at para. 107; see Professor Hamish Stewart, Fundamental Justice: Section 7 of the Canadian Charter of Rights and Freedoms (Toronto: Irwin Law Inc., 2019), at pp. 150, 187.

[207] Professor Stewart points out that each of these principles is not a silo operating entirely independently; they are connected: Fundamental Justice, at p. 189; Bedford, at para. 109; Canada (Attorney General) v. PHS Community Services Society, [2011] 3 S.C.R. 134, [2011] S.C.J. No. 44 2011 SCC 44, at para. 127. Each case usually responds more aptly to one principle, while the other principles offer different perspectives. Moreover, "it is possible for a law to offend one of these norms without offending the other two": Fundamental Justice, at p. 189.

[208] The Supreme Court identified two "evils" at which these principles are directed. The "first evil" is the "absence of connection between the law's purpose and the s. 7 deprivation", which engages the principles of arbitrariness and overbreadth: Bedford, at para. 108. The "second evil" arises where the law's effects on an individual's life, liberty, or security of person operate "in a manner that is grossly disproportionate to the law's objective", and therefore engages the third principle of gross disproportionality: Bedford, at para. 109.

[209] Overbreadth is engaged when a law is so broad it captures some conduct that bears no rational connection to its purpose: Bedford, at para. 112. This principle recognizes that the law may be "rational in some cases, but that it overreaches in its effect in others": Bedford, at para. 113. Professor Stewart describes overbreadth as the "dominant" principle of the three: Fundamental Justice, at p. 152.

[210] The principle of "arbitrariness" exists where there is no "direct" or "rational" connection "between the purpose of the law and the impugned effect on the individual", or if it can be shown that the impugned effect undermines the law's objective: [page406] Bedford, at para. 111. In the area of any overreach, the law is to be understood as arbitrary. That is why the principles of arbitrariness and overbreadth are related but distinct principles: Bedford, at para. 117.

[211] Gross disproportionality considers whether "the law's effects on life, liberty or security of the person" are so disproportionate that "the deprivation is totally out of sync with the objective of the measure": Bedford, at para. 120.

[212] If a law violates one of these principles, "there is a mismatch between the legislature's objective and the means chosen to achieve it": Fundamental Justice, at p. 150.

[213] Methodologically, it is necessary to first identify the objective of s. 33.1 in order to execute the s. 7 analysis. The trial judge identified the law's objectives in his s. 1 analysis, which will serve the purpose here. He found that the objectives of s. 33.1 are those stated in the Preamble to Bill C-72 and are: "the protection of women and children from intoxicated violence and ensuring the accountability of those who commit offences of violence while intoxicated": at paras. 115, 121.

[214] I agree that Parliament was seeking to discourage "self-induced intoxication," which it described as "blameworthy", in order to prevent violence, "particularly violence against women and children", for which persons "should be held accountable": see Preamble to Bill C-72. As noted, there is a high correlation between self-induced alcohol intoxication and violence, particularly violence by intoxicated men against women and children. This is the "protective objective" of s. 33.1.

[215] The trial judge referred to "accountability" as another objective of s. 33.1. I interpret his use of "accountability" as intending to capture the penal objective of s. 33.1. In submissions to Parliament prior to the enactment of Bill C-72, there was a pervasive sense of outrage at the prospect that a person who sexually assaulted an elderly disabled woman might be permitted to escape punishment on the ground of excessive intoxication, as was the case in Daviault. This reflects a deep intuition of justice that those who commit such terrible acts should be equally subject to penal consequences and should pay the same price, excessively intoxicated or not. Such acts should never be consequence free. More precisely, those who could have sheltered under the defence of non-mental disorder automatism are now, under s. 33.1, subjected to the same penal consequences for their violent acts as those whose state of intoxication was slightly less so as not to be in a state of automatism when committing those same acts. This [page407] is the "penal objective" of s. 33.1. The protective and the penal objectives are related but also separate and distinct.

[216] Although it is plausible that the legislation could discourage people from extreme alcohol intoxication, that dynamic would not apply to people like Mr. Chan and Mr. Sullivan. Neither was drinking. Neither had any reason to believe that their voluntary self-intoxication would culminate in violent psychosis. Common sense suggests that s. 33.1 would not discourage people who lack any basis for believing that self-intoxication would cause them to become psychotic from self-intoxicating. Because their conduct was captured under s. 33.1, the provision is overbroad in the Bedford sense because there is no connection between the law's objectives and its effects on the appellants: Bedford, at para. 112. The law is also arbitrary in the area of overbreadth because its effects on Mr. Chan and Mr. Sullivan bear no connection to its stated objectives: it punishes those who did not foresee that self-intoxication would lead to acts of violence.

[217] An overbroad law adversely impacting an individual's s. 7 rights is sufficient to establish a limit. The Supreme Court stated in Bedford: "The question under s. 7 is whether anyone's life, liberty or security of the person has been denied by a law that is inherently bad; a grossly disproportionate, overbroad, or arbitrary effect on one person is sufficient to establish a breach of s. 7" (emphasis in original): at para. 123. It is plain that the s. 7 rights of both Mr. Chan and Mr. Sullivan have been limited by s. 33.1 of the Criminal Code, along with other similarly situated persons.

[218] The s. 7 analysis in these appeals proceeds at two levels. The first is the assessment undertaken by the trial judge in the more traditional pattern of Daviault. It identified several limits resulting from bypassing the voluntariness and mental element requirements and the presumption of innocence for criminal convictions on the predicate violence-based charges. These limits apply generally to all those to whom s. 33.1 applies and are about as fundamental as rights get in the criminal context.

[219] The application of the Bedford/Carter framework adds an element. It shows that s. 33.1 is overbroad and arbitrary in its application specifically to Mr. Chan and Mr. Sullivan. Their s. 7 rights were limited by the operation of s. 33.1 in the sense that there is no connection between the law's two objectives, protective and penal, and the law's effects on the appellants.

[220] The next question is whether the identified limits can be demonstrably justified under s. 1 of the Charter. [page408]

[221] Section 33.1 of the Criminal Code limits the Charter rights of the appellants under ss. 7 and 11(d). If the Crown fails to discharge its burden under s. 1 of demonstrating that s. 33.1 imposes "reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society", then s. 52(1) of the Constitution Act, 1982 will deem s. 33.1 to be "of no force or effect" to the extent of any inconsistency with the Charter.

[222] I need not restate the Oakes framework: R. v. Oakes, 1986 CanLII 46 (SCC), [1986] 1 S.C.R. 103, [1986] S.C.J. No. 7.

[223] I begin by addressing whether the s. 1 analysis in Daviault is dispositive of these appeals. From there, I continue with observations about the contextual approach to the s. 1 analysis; the role of judicial deference to Parliament; and the current approach to the Oakes analysis. The interplay of ss. 7 and 1 in the proportionality analysis makes these appeals somewhat unusual.

[224] While Daviault provides guidance, it is not dispositive of the s. 1 analysis. As with any legal doctrine, earlier holdings are subject to modification by later doctrinal developments.

[225] The Daviault majority found that similar rights limitations established by the common law in Leary could not be justified under s. 1 of the Charter. Cory J. explained, "this rare and limited defence in general intent offences is required so that the common law principles of intoxication can comply with the Charter": at para. 67. This conclusion flowed from his earlier statement that: "the Leary rule applies to all crimes of general intent, it cannot be said to be well tailored to address a particular objective and it would not meet either the proportionality or the minimum impairment requirements": at para. 47. Cory J. added that, because there was an insufficient link between intoxication and criminal acts and, "under the proposed approach, the defence would be available only in the rarest of cases . . . there is no urgent policy or pressing objective which needs to be addressed": at para. 47.

[226] However, the Daviault majority held that the reverse onus created by placing the burden on the accused to establish automatism on the balance of probabilities, with the assistance of expert evidence, was a reasonable limit of the accused's Charter [page409] rights: at para. 63. The reverse onus has not been raised as an issue in these appeals.

[227] Since 1994, when Daviault was decided, the doctrine has evolved with experience and is considerably more nuanced, particularly as the result of Alberta v. Hutterian Brethren of Wilson Colony, [2009] 2 S.C.R. 567, 2009 SCC 37, Bedford, and Carter. Further, Parliament's legislative response to Daviault must be assessed with fresh judicial eyes: R. v. Mills, 1999 CanLII 637 (SCC), [1999] 3 S.C.R. 668, [1999] S.C.J. No. 68.

[228] It is now trite law that the s. 1 analysis is contextual and fact-specific. McLachlin J. (as she then was) observed that "the Oakes test must be applied flexibly, having regard to the factual and social context of each case": RJR-MacDonald Inc. v. Canada (Attorney General), 1995 CanLII 64 (SCC), [1995] 3 S.C.R. 199, [1995] S.C.J. No. 68, at para. 132. The court's proper role "will vary according to the right at issue and the context of each case" and "cannot be reduced to a simple test or formula": Doucet-Boudreau v. Nova Scotia (Minister of Education), [2003] 3 S.C.R. 3, [2003] S.C.J. No. 63, 2003 SCC 62, at para. 36.

[229] There are several considerations that bear on the legitimacy of judicial review of legislation and on how that power should be exercised by courts.

[230] The first is the separation of powers and the development of "certain core competencies in the various institutions": Ontario v. Criminal Lawyers' Association of Ontario, [2013] 3 S.C.R. 3, [2013] S.C.J. No. 43, 2013 SCC 43 ("CLAO"), at para. 28, per Karakatsanis J.

[231] The branches of government, with their different institutional capacities, "play critical and complementary roles in our constitutional democracy . . . [which] each branch will be unable to fulfill . . . if it is unduly interfered with by the others": CLAO, at para. 29. This requires the judiciary to be deferential not only to policy objectives, but also to the specific means Parliament chooses to achieve those objectives: Doucet-Boudreau, at para. 57.

[232] The court must respect the core competencies to which Karakatsanis J. referred in CLAO. Criminalizing socially harmful conduct is a core competency conferred on Parliament: Constitution Act, 1867, 30 & 31 Vict., c. 3, s. 91. This is where the democratic principle has its greatest force. The Criminal Code embodies Parliament's primacy in creating criminal offences; the [page410] court is prohibited from creating both common law criminal offences and new common law defences that would be inconsistent with the Code's provisions: ss. 8(3), 9.

[233] The core competency of Parliament over the criminal law is implicated deeply in these appeals. In assessing the constitutionality of legislation, the court must be mindful that "in certain types of decisions there may be no obviously correct or obviously wrong solution, but a range of options each with its advantages and disadvantages": Newfoundland (Treasury Board) v. Newfoundland & Labrador Association of Public and Private Employees, [2004] 3 S.C.R. 381, [2004] S.C.J. No. 61, 2004 SCC 66, at para. 83, per Binnie J. While judicial deference is due, it never amounts to submission because that would abrogate the court's constitutional responsibility: Gordon v. Canada (Attorney General), [2016] O.J. No. 4330, 2016 ONCA 625, at para. 236, leave to appeal refused [2016] S.C.C.A. No. 444, [2016] S.C.C.A. No. 445; see also P.S.A.C. v. Canada, 1987 CanLII 89 (SCC), [1987] 1 S.C.R. 424, [1987] S.C.J. No. 9, at para. 36; RJR-MacDonald, at para. 136.

[234] The Oakes framework is intended to structure the legal analysis and thereby to constrain and discipline courts, in order to render the final balancing step as intelligible and as transparent as possible.

[235] The Oakes test was not mandated by s. 1. It was developed by the Supreme Court as a means of structuring the inquiry into whether a limit on the exercise of a Charter right is demonstrably justified in a free and democratic society. A flexible, contextual approach has won out over a rigid application of Oakes, especially in the wake of Hutterian Brethren, Bedford, and Carter; see also Gérard La Forest, "The Balancing of Interests under the Charter" (1992), 2 N.J.C.L 133, at pp. 145-148. In La Forest J.'s view, the Oakes test is not a set of rigid rules, but "a checklist, guidelines for the performance" of judicial duties: at 148.

[236] The Bedford/Carter framework for assessing s. 7 limits introduces a new dynamic into the s. 1 proportionality analysis. The Supreme Court prescribed a division of labour between ss. 7 and 1. As noted earlier, the s. 7 analysis addresses the impact on the individual in isolation from society. It is only in the s. 1 analysis that the common good is considered: Bedford, at paras. 121, 124-29. [page411]

[237] Given this analytical division, it is important to stress that a finding that a s. 7 right has been limited is not determinative of the s. 1 analysis.

[238] In Bedford, the Supreme Court did not discount the prospect that a s. 7 limit could be justified under s. 1, despite statements in earlier cases that considered this to be unlikely given the "significance of the fundamental rights protected by s. 7": at para. 129. The court did not undertake the s. 1 analysis in Bedford, but did so in Carter, and agreed with the trial judge that the absolute prohibition on physician-assisted dying was overbroad in s. 7 terms and was not minimally impairing in s. 1 terms: Carter, at paras. 86, 88 and 121.

[239] The individualized application of the Bedford/ Carter framework to Mr. Chan and Mr. Sullivan found s. 33.1 to be overbroad. Where does a finding of overbreadth under s. 7 fit into the s. 1 analysis? The question has not yet been answered definitively. In Heywood, Cory J. said that it fits best into the minimal impairment step: R. v. Heywood, 1994 CanLII 34 (SCC), [1994] 3 S.C.R. 761, [1994] S.C.J. No. 101, at para. 69. The Supreme Court took the same approach in Carter: at paras. 102-21.

[240] In my view, although it provides additional conceptual tools, the Bedford/Carter framework does not displace, but rather supplements, the traditional approach taken in Daviault and by the trial judge in Chan.

[241] In the s. 1 analysis, the Crown must first establish that the legislation is "in pursuit of a sufficiently important objective that is consistent with the values of a free and democratic society" and is of "sufficient importance to warrant overriding a constitutionally protected right or freedom": R. v. J. (K.R.), 2016 SCC 31 (CanLII), [2016] 1 S.C.R. 906, [2016] S.C.J. No. 31, at para. 61; Oakes, at para. 69.

[242] Determining whether an objective is pressing and substantial is usually not an evidentiary contest. As the Supreme Court has explained: "The proper question at this stage of the analysis is whether the Attorney General has asserted a pressing and substantial objective"; even a "theoretical objective asserted as pressing and substantial is sufficient for purposes of the s. 1 justification analysis": Harper v. Canada (Attorney General), [2004] 1 S.C.R. 827, [2004] S.C.J. No. 28, 2004 SCC 33, at paras. 25-26. [page412]

[243] As noted earlier, the trial judge identified the objectives of s. 33.1 as those stated in the Preamble to Bill C-72: "the protection of women and children from intoxicated violence" and "the accountability of those who commit offences of violence while intoxicated": at para. 121.

[244] In my view, this is not a case in which there is a live issue about how the legislative objectives are to be identified, as in Bedford, Carter, and Frank v. Canada (Attorney General), [2019] 1 S.C.R. 3, [2019] S.C.J. No. 1, 2019 SCC 1. Nor do the appellants argue that Parliament proceeded in bad faith or had an unexpressed ulterior motive, as was alleged in PSAC.

[245] There is a debate on whether the objectives stated in the Preamble can be taken at face value, or whether the real objectives of s. 33.1 lie elsewhere. On the one hand, the appellants argue that the trial judge in Chan overstated the objective -- s. 33.1 was designed to abolish the Daviault defence in the wake of public pressure. On the other hand, the Crown argues that such a narrow interpretation of the objective implies bad faith and undermines the presumption that Parliament intends to enact constitutional legislation: see Mills, at paras. 48, 56-60.

[246] I find arguments that the objectives are either too broadly or too narrowly framed in the Preamble not to be compelling.

[247] What was Parliament trying to do by enacting s. 33.1? In its own words, Parliament was seeking to discourage "self-induced intoxication", which it described as "blameworthy", in order to prevent violence, "particularly violence against women and children", for which persons "should be held . . . accountable": Preamble to Bill C-72.

[248] It is noteworthy that Canada is not alone in enacting such legislation. The United Kingdom, Australia, and the United States have enacted similar laws that go further in their application. For instance, in the United Kingdom, voluntary intoxication is never a defence for general intent offences. Compared to those jurisdictions, Canada's approach is narrower: the application of s. 33.1 is limited to violent general intent offences.

[249] Some intervenors argued that whether an objective is pressing and substantial depends in part on the incidence of the application of the limiting measures. If something is rarely applied, goes the argument, the objective cannot be pressing and substantial. This seems to have been Cory J.'s thinking in Daviault, [page413] when he said that because: "the defence would be available only in the rarest of cases . . . there is no urgent policy or pressing objective which need[s] to be addressed": at para. 47.

[250] I would reject this argument. The determination of whether a matter is of sufficient importance to the public good does not turn on statistical frequency. There are several criminal offences whose commission is quite rare, like treason or espionage, but no one would argue that Parliament's objectives in criminalizing such conduct are not pressing and substantial. Rarity of occurrence, in itself, does not impugn the pressing and substantial nature of Parliament's objectives.

[251] It bears repeating that the focus at this step of the analysis is whether the Crown has "asserted" a pressing and substantial objective, even a theoretical one: Harper, at paras. 25-26. From a democratic viewpoint, the court should presume that Parliament "intended to enact constitutional legislation and strive, where possible, to give effect to this intention": Mills, at para. 56.

[252] In my view, Parliament has answered the question at this stage of the analysis -- its protective and penal objectives in enacting s. 33.1 are self-evidently pressing and substantial objectives.

[253] The next stage of the analysis asks if the methods Parliament chose to achieve these objectives are proportional.

[254] The task in the second stage of Oakes is to make the proportionality determination. As noted, the court must assess whether the means chosen by Parliament to accomplish its ends are: first, rationally connected with the ends; second, minimally impairing; and third, proportional as between the deleterious and salutary effects of the law: Oakes, at para. 70; Carter, at para. 94; J. (K.R.), at para. 58; Frank, at paras. 38-39. The measures identified in these reasons as limiting the appellants' Charter rights are the means to be assessed against Parliament's ends in determining whether it is just for Parliament to require some individuals to bear the negative effects of the measures in order to secure the positive effects of the ends for the common good.

[255] To recapitulate, s. 33.1 removes the common law defence of non-mental disorder automatism created by the Supreme Court in Daviault. Now, those who could have sheltered under the defence of non-mental disorder automatism are subjected to the same penal consequences for their violent acts as those who commit the same acts while in a less intoxicated, non-automatistic state. [page414]

[256] The evidentiary burden at this stage of the proportionality analysis "is not particularly onerous", and as Professor Peter Hogg commented, "the requirement of a rational connection has very little work to do": Little Sisters Book and Art Emporium v. Canada (Minister of Justice), [2000] 2 S.C.R. 1120, [2000] S.C.J. No. 66, 2000 SCC 69, at para. 228; Peter Hogg, Constitutional Law of Canada, 5th ed., vol. 2 (Toronto: Carswell, 2019) (looseleaf updated 2019), at s. 38.12.

[257] A rational connection need not be proven on a rigorous scientific basis. A causal connection based on reason or logic may suffice: RJR-MacDonald, at paras. 137, 153 and 156; Carter, at para. 99. Provided that the impugned measure shows care in design and a lack of arbitrariness, and provided that it furthers an important government aim in a general way, it will pass the rational connection branch of the analysis: Canada (Human Rights Commission) v. Taylor, 1990 CanLII 26 (SCC), [1990] 3 S.C.R. 892, [1990] S.C.J. No. 129, at para. 56.

[258] The Crown need only demonstrate a reasonable prospect that the limiting measure will further the objective to some extent, not that it will certainly do so: Hutterian Brethren, at para. 48. In the absence of dispositive social science evidence, Parliament need only establish a "reasoned apprehension" of the harm it aims to prevent: R. v. Butler, 1992 CanLII 124 (SCC), [1992] 1 S.C.R. 452, [1992] S.C.J. No. 15, at para. 107. For example, in McKinney v. University of Guelph, some evidence showed a correlation between mandatory retirement and generating new jobs for younger faculty, while other evidence suggested that there was none: 1990 CanLII 60 (SCC), [1990] 3 S.C.R. 229, [1990] S.C.J. No. 122, at para. 65. The court found that this conflicting evidence provided a sufficient basis to meet the rational connection test: at para. 66.

[259] Where the legislation at issue has more than one objective, any of them can be relied upon to meet the s. 1 test: Hutterian Brethren, at paras. 44-45.

[260] The trial judge in Chan expressed the view that s. 33.1 does not do much to protect women and children from violence. He said: "I have a hard time believing, as a matter of common sense, that many individuals are deterred from drinking, in the off chance that they render themselves automatons and hurt someone": at para. 123. While he accepted the evidence that: "there is a strong linkage between intoxication and violence", he [page415] stated that because: "the self-induced automatism defence arises very rarely and is successful even more rarely . . . I am unable to conclude what the actual connection is between the objective and what the law will actually achieve in terms of reducing violence against women and children": at paras. 125-26. While he found that the limiting measures in s. 33.1 were not rationally connected to the protective objective, he concluded that they were rationally connected to the penal objective: at paras. 126-27.

[261] The Preamble to Bill C-72 (the legislation enacting s. 33.1) expressly points to the "close association between violence and intoxication". That correlation is well-established in the evidence considered by Parliament in its deliberations on s. 33.1, but it falls short of showing that those who become intoxicated intend to commit assaults. Correlation -- a statistical concept -- and causation, which is essential to criminal liability, are quite different. It is not obvious that substantial correlation cannot form the basis of a legislative response.

[262] The second issue was whether excessive alcohol intoxication can physiologically lead to non-mental disorder automatism. The Ontario Crown's factum again fairly summarizes the expert evidence:

[263] In my view, the social science evidence on these two points, even though not dispositive, does establish a "reasoned apprehension" capable of grounding s. 33.1 as a rational social policy response by Parliament to a real problem. The Preamble reflects both these issues and expressly refers to the social science evidence.

[264] Does this evidence support both the protective objective and the penal objective of s. 33.1?

[265] I would agree with the trial judge that the introduction of the non-mental disorder automatism Daviault defence is unlikely, as a matter of logic and common sense, to have encouraged any excessive drinkers, in the moment, to cross the brink into automatism in order to get access to the new defence. Nor is it likely that the removal of the defence, conversely, will encourage excessive drinkers, in the moment, to drink less. Furthermore, there is no evidence before this court to support the Crown's argument that the measure will discourage excessive drinking. On this basis, the trial judge correctly found that s. 33.1's measures are not rationally connected to the protective objective.

[266] However, the penal objective is not merely ancillary to the protective objective. It has independent status, in view of the public outrage that greeted the Daviault decision, the social science evidence, and the submissions to Parliament that demanded full criminal liability for those who, having committed a violent assault, would be able to shelter under the new non-insane automatism defence. The measures in s. 33.1 are rationally connected to the penal objective.

[267] More recently, the Supreme Court outlined this step of the Oakes test in Carter and explained that "the analysis at this stage is meant to ensure that the deprivation of Charter rights is confined to what is reasonably necessary to achieve the state's object": at para. 102. The question at this stage, therefore, is whether "the limit on the right is reasonable tailored to the objective . . . [in particular] 'whether there are less harmful means of achieving the legislative goal'" (internal citations omitted): at para. 102. As such, the government bears the burden of demonstrating that "less drastic means" were unavailable to "[achieve] the objective 'in a real and substantial manner'": at para. 102. [page417]

[268] Judicial deference to Parliament at the minimal impairment stage is sensitive to the context of the law in issue. The Supreme Court has affirmed that Parliament is not held to a standard of perfection: "If the law falls within a range of reasonable alternatives, the courts will not find it overbroad merely because they can conceive of an alternative which might better tailor objective to infringement": Libman v. Quebec (Attorney General), 1997 CanLII 326 (SCC), [1997] 3 S.C.R. 569, [1997] S.C.J. No. 85, at para. 58, citing McLachlin J.'s formulation of the test in RJR-MacDonald, at para. 160; see also Montreal (City) v. 2952-1366 Quebec Inc., [2005] 3 S.C.R. 141, [2005] S.C.J. No. 63, 2005 SCC 62, at para. 94. Rather, the court will consider whether the government has established that "it has tailored the limit to the exigencies of the problem in a reasonable way": Montreal (City), at para. 94; see also Nova Scotia (Workers' Compensation Board) v. Martin, [2003] 2 S.C.R. 504, [2003] S.C.J. No. 54, 2003 SCC 54, at para. 112, per Gonthier J.; Lavigne v. O.P.S.E.U., 1991 CanLII 68 (SCC), [1991] 2 S.C.R. 211, [1991] S.C.J. No. 52, at para. 170, per Wilson J.

[269] It is not always the case that an overbroad law will automatically fail at the minimal impairment stage: Fundamental Justice, at pp. 364-365; see also Bedford, at para. 144; Carter, at paras. 102-21; Michaud, at paras. 73-74. As Professor Stewart explains, "under section 1, the issue is whether the limit on the right impairs the section 7 right no more than reasonably necessary to achieve the purpose of that limit" (emphasis added): Fundamental Justice, at p. 364.

[270] The trial judge found that s. 33.1 was minimally impairing because its limits were tailored to its objectives and the provision fell "within a range of reasonable alternatives": at paras. 140-141, citing RJR-MacDonald, at para. 160, per McLachlin J.

[271] While he recognized that the law impaired Mr. Chan's rights in a "certainly not minimal" manner, the trial judge noted that s. 33.1 has three mitigating features: the section only applies to general intent offences, not the more serious specific intent offences; it only applies to offences relating to bodily integrity and not to property-based offences; and it only applies to voluntary self-induced intoxication: at paras. 131-33, 141. In his view: "There is a moral blameworthiness attached to getting oneself so intoxicated as to lose control of one's faculties": at para. 134.

[272] The trial judge did not accept that creating a stand-alone offence of criminal intoxication was a reasonable alternative to the measures in s. 33.1. He noted that Parliament had rejected this option on the reasoned basis that: there might be seen to be a [page418] "discount" available for some intoxicated offenders in the form of a reduced sentence; it would undermine the objective of accountability, or, as I have framed the point, it would not advance Parliament's penal objective; and it would require prosecutors to argue somewhat inconsistently that an accused was not so intoxicated to avoid responsibility for the predicate offence, but was sufficiently intoxicated to be guilty of criminal intoxication: at para. 139.

[273] There is no doubt, on the evidence presented, that Parliament wanted to achieve the penal objective: to subject those who could have sheltered under the defence of non-mental disorder automatism to the same penal consequences for their violent acts as those who commit the same acts in less intoxicated, non-automatistic states. This is what must run the gauntlet at the balancing step of the proportionality assessment, giving due weight to Parliament's authority to criminalize socially harmful conduct.

[274] The s. 7 finding of overbreadth must now be considered. The application of the Bedford/Carter framework shows s. 33.1 does not limit s. 7 rights in a minimally impairing manner. It is overbroad as applied to Mr. Chan and Mr. Sullivan. Their s. 7 rights are limited by the application of s. 33.1 because there is no connection between the law's two objectives, protective and penal, and the law's effects on them. Neither Mr. Chan nor Mr. Sullivan was drinking. Neither had any reason to believe that his voluntary self-intoxication would culminate in violent psychosis.

[275] These observations apply to similarly situated individuals who have no reason to believe that their voluntary self-intoxication would culminate in violent psychosis. For instance, similarly situated individuals who take prescription drugs and experience unanticipated side effects, or people who voluntarily consume intoxicants other than alcohol and could not foresee that doing so would lead to violent psychosis, are captured by the law, according to the Crown. By attaching criminal liability to involuntary conduct or situations where an individual consumes a drug in circumstances where violent psychosis is not reasonably foreseeable, s. 33.1 creates a standard of absolute liability. Moreover, common sense suggests that s. 33.1 would not discourage people who lack any reasonable basis for believing that self-intoxication would cause them to become psychotic from becoming intoxicated. While it is open to Parliament to craft an offence for committing "a prohibited act while drunk", the means employed would not be minimally impairing if they bear no connection to the law's objectives: Daviault, at para. 61. [page419]

[276] Parliament took pains to tailor s. 33.1 to its stated objectives, as noted by the trial judge. In my view, the result of those efforts falls within the range of reasonable alternatives, since: "The tailoring process seldom admits of perfection and the courts must accord some leeway to the legislator": Libman, at para. 58; RJR-MacDonald, at para. 160.

[277] However, because s. 33.1 does not contain an exception for people like Mr. Chan and Mr. Sullivan and those similarly situated, it is not minimally impairing. But s. 33.1 still must be evaluated for proportionality under s. 1, to which I now turn.

[278] In Hutterian Brethren, McLachlin C.J.C. set out the basic question at the third step of the proportionality analysis: "Is the limit on the right proportionate in effect to the public benefit conferred by the limit?": at para. 73. Or, as she put it in Bedford, "whether the negative impact of a law on the rights of individuals is proportionate to the pressing and substantial goal of the law in furthering the public interest": at para. 125; see also Carter, at para. 95. This analysis "takes full account of the 'severity of the deleterious effects of a measure on individuals or groups'": Hutterian Brethren, at para. 76. It entails a broad assessment of whether the "benefits of the impugned law are worth the cost of the rights limitation", or whether "the deleterious effects are out of proportion to the public good achieved by the infringing measure": at paras. 77-78.

[279] The trial judge noted, adverting to Bedford, that the broader social goals of s. 33.1 were to be taken into account in the balancing of rights: at para. 150. He noted: "The entire history of the defence of intoxication has been about finding the appropriate balance between the fundamental rights of accused persons and the rights of others -- particularly women and children -- to be protected from intoxication-fueled violence": at para. 153. He deferred to Parliament's "view of the morally appropriate balance between intoxicated offenders and the rest of society and to hold intoxicated offenders to account": at para. 154. The trial judge concluded that he was "satisfied that there is proportionality between the salutary and deleterious effects of the provision": at para. 157. [page420]

[280] In the balancing step, the court is required to consider proportionality as between the deleterious and salutary effects of the law, in order to determine whether it is just for the legislation to require some individuals to bear the burden of the negative effects in order to secure the benefits of the positive effects for the common good -- a good that benefits the offender as well: Oakes, at para. 70; Carter, at para. 94; J. (K.R.), at para. 58; Frank, at paras. 38-39.

[281] The parties supporting the constitutionality of s. 33.1 identified several benefits that inure to the common good. First, the section satisfies the deep-seated conviction that it would be wrong and unjust to allow a person like Mr. Daviault, who committed a violent sexual assault on a disabled elderly woman while voluntarily self-intoxicated, to escape penal consequences. This conviction drove the submissions to Parliament, and Parliament itself in enacting the legislation. It is not a conviction to be taken lightly in a free and democratic society.

[282] Second, s. 33.1 subjects all those who voluntarily self-intoxicate and then commit violent assaults to the same penalties, which sends a strong message of deterrence to the public that this conduct will not be tolerated.

[283] Third, LEAF asserts that protecting the security of the person and equality rights of others, particularly women and children, from violent crimes at the hands of intoxicated offenders depends on the reporting of violent abuse by victims and witnesses. As the Minister of Justice explained in Parliamentary debate, the uncertainty left by Daviault would discourage victims and witnesses from reporting such drunken assault. Removing the defence would, it is argued, remove a disincentive to report.

[284] Fourth, these benefits all inure to the special benefit of women and children who are the primary victims of intoxicated violence, and who have been recognized in many legislative enactments as being vulnerable and requiring legal protection.

[285] Fifth, s. 33.1 places the fault where it belongs -- with those who would voluntarily self-intoxicate to excess, which is not a morally blameless act.

[286] All of these benefits are not ex post facto rationalizations but appear in some form in the Preamble to Bill C-72.

[287] The countervailing burdens are weighty. The fundamental rights of persons caught by s. 33.1 under ss. 7 and 11(d) of the [page421] Charter are severely limited: these are, to repeat, the presumption of innocence and the strong criminal law requirement that the Crown prove beyond a reasonable doubt that the violent acts of the accused were voluntary and met the mental element requirements for criminal convictions on the predicate violence-based charges.

[288] While it is true that the incidence of the application of s. 33.1 is rare, that does not justify depriving even such a small number of persons of their fundamental rights. The dramatic effect on these rights is disproportionate to the small number of individuals affected. Further, Parliament's core target under s. 33.1 was the person whose extreme alcohol intoxication would cause non-mental disorder automatism. But it is not clear that extreme alcohol intoxication causes non-mental disorder automatism as a matter of basic science. In short, the defence might not even be viable as a matter of fact. (Mr. Daviault was not tried again because his victim died before the second trial of unrelated causes.)

[289] The final step of the proportionality analysis turns on a conviction sedimented deeply into the rule of law. The principle that "the innocent [should] not be punished" has been recognized "from time immemorial [as] part of our system of laws", a system "founded upon a belief in the dignity and worth of the human person and the rule of law": Reference re Motor Vehicle Act (British Columbia) Section 94(2), 1985 CanLII 81 (SCC), [1985] 2 S.C.R. 486, [1985] S.C.J. No. 73, at p. 513 S.C.R., per Lamer J. The Supreme Court reiterated a variant of this conviction in Carter, explaining that "a law that runs afoul of the principles of fundamental justice" is not easily justified or "overridden by competing societal interests" (internal citations omitted): at para. 95.

[290] I share the conviction stated in these authorities. Section 33.1 cannot be justified under s. 1 of the Charter, as a matter of simple justice and what our law requires the Crown to prove in order to secure a criminal conviction for the predicate violent offences. The required mental and voluntariness elements and the presumption of innocence cannot be bypassed.

[291] Finally, focusing specifically on Mr. Chan and Mr. Sullivan, there is no good reason for them to have been swept into the net of s. 33.1. Section 33.1 is overbroad in its application to them because there is no connection between the law's two objectives -- protective and penal -- and the law's effects on them. The justification of "enforcement practicality" does not apply to them, nor does the phenomenon of line-drawing apply, as in Michaud, which can give rise to incidental overbreadth and arbitrariness: [page422] Bedford, at para. 113; Michaud, at paras. 144-45. Section 33.1 imposes an even more intense limit on their rights than the general operation of s. 33.1 and attracts the same evaluation of unconstitutionality.

[292] For these reasons, in my view, s. 33.1 of the Criminal Code limits the Charter rights of the appellants under ss. 7 and 11(d) and the Crown has not demonstrated that those limits are justified. Consequently, s. 52(1) of the Constitution Act, 1982 deems s. 33.1 to be "of no force or effect" to the extent of any inconsistency with the Charter. I therefore concur with my colleague in the result.