"Sexsomniac" cleared of rape charge - Page 31

E. ANALYSIS
[52] Before testing the trial judge’s characterization of the respondent’s parasomnia as non-mental disorder automatism against the binding jurisprudence, I will sketch in the necessary legal background.
(i) Automatism Claims

[53] Conduct that is not voluntary cannot be criminal: Rabey, per Ritchie J. (majority) at p. 6, per Dickson J. (dissenting on other grounds) at p. 26; Parks.[3] The voluntariness requirement is a principle of fundamental justice protected by s. 7 and s. 11(d) of the Charter: R. v. Daviault 1994 CanLII 61 (S.C.C.), (1994), 93 C.C.C. (3d) 21 (S.C.C.), at pp. 48-49, 69.

[54] A claim by an accused that his or her conduct was involuntary and should result in an acquittal for that reason can arise in a variety of very different circumstances. Automatism is the legal term used to describe one specific kind of involuntary action: see Bratty v. Attorney General for Northern Ireland, (1963) 3 All E.R. 523 (H.L.), per Lord Denning at pp. 408-409; Parks, per La Forest J. at p. 302. Automatism refers to involuntary conduct that is the product of a mental state in which the conscious mind is disassociated from the part of the mind that controls action. A person in a state of automatism may perform acts, sometimes complicated and apparently purposeful acts, but have no control over those actions: William Wilson, et al., “Violence, Sleepwalking and the Criminal Law: (2) The Legal Aspects” (2005) Crim. L.R. 614, at pp. 615-16. North P. put it this way in R. v. Burr, [1969] N.Z.L.R. 736 (C.A.), at p. 744:
In my opinion then there is now clear judicial authority for the view that in order for a defence of automatism to succeed, the person whose conduct is under review must be unconscious of what he was doing. In short that what he did was an unconscious involuntary act… [I]n my opinion, the evidence must be sufficient to lay a proper foundation for the plea that the accused person acted through his body and without the assistance of his mind, in the sense that he was not able to make the necessary decisions and to determine whether or not to do the act. [Emphasis added.]

[55] The disassociative state that is the hallmark of automatism can be caused by many things including disease, mental illness, concussion, drugs, and parasomnia. Each of these conditions can produce a condition in which an accused, while capable of complex, apparently goal-oriented conduct, is incapable of exercising any control over those actions. As will be discussed below, the cause of the automatism is an important consideration in characterizing the nature of the automatism.

[56] The automatism “defence” is not a defence in the true sense but is a denial of the commission of the actus reus of the crime. Absent the commission of the prohibited act, there can be no crime and hence no criminal liability. A person who is unable to decide whether to perform an act and unable to control the performance of the act cannot be said, in any meaningful sense, to have committed the act. Nor can it be appropriate in a criminal justice system in which liability is predicated on personal responsibility to convict persons based on conduct which those persons have no ability to control: see Stone, per Bastarache J. at p. 417, per Binnie J. (dissent) at pp. 378-79; Don Stuart, Canadian Criminal Law: A Treatise, 5th ed. (Toronto: Carswell, 2007), at pp. 107-110; David Ormerod, Smith & Hogan Criminal Law, 11th ed. (New York: Oxford University Press, 2005), at pp. 44-46; Andrew Ashworth, Principles of Criminal Law, 5th ed. (New York: Oxford University Press, 2006), at pp. 98-100; Stanley Yeo, “Putting Voluntariness Back Into Automatism” (2001) 32 V.U.W.L.R. 15.

[57] Automatism claims raise legitimate questions about an accused’s mental status and his or her potential danger to the public. These claims are almost always advanced in cases where the accused has caused serious personal injury or at least put members of the public at serious risk. It hardly seems controversial that persons who engage in what would otherwise be regarded as serious criminal conduct and claim to have had absolutely no control over that conduct should have their mental health and their right to remain at liberty scrutinized. Automatism claims, which by their very nature assert that the accused acted while in an abnormal and impaired mental state, inevitably bring into play the exemption to criminal liability created by s. 16 of the Criminal Code.

[58] Section 16 exempts persons from criminal responsibility if, as a result of a “mental disorder”, they were incapable of appreciating the nature and quality of their acts or of knowing that their acts were wrong. As Binnie J. observed in dissent in Stone, at paras. 78-83, s. 16 does not speak to the voluntariness of one’s actions but instead addresses cognitive functions that assume voluntary conduct. It is, however, accepted in Canada, and throughout the Commonwealth, that a person whose conduct is involuntary because of a condition that is the product of a mental disorder falls exclusively within the purview of the “insanity” defence: see R. v. Revelle 1979 CanLII 79 (ON C.A.), (1979), 48 C.C.C. (2d) 267 (Ont. C.A.), at p. 271, aff’d 1981 CanLII 48 (S.C.C.), (1981), 61 C.C.C. (2d) 575; Rabey (Ont. C.A.), at p. 472, (S.C.C.), at p. 6; Stone, per Bastarache J., for the majority, at paras. 160-161; Attorney-General’s Reference (No. 2 of 1992), [1994] Q.B. 91 (C.A.), at pp. 104-105; R. v. Falconer (1990), 171 C.L.R. 30 (H.C.); The Queen v. Cottle, [1958] N.Z.L.R. 999 (C.A.).

[59] If an accused’s automatism is rooted in a mental disorder, the accused will not be acquitted but will be found NCR-MD. Under Part XX.I of the Criminal Code, a person found NCR-MD is subject to a post-verdict disposition hearing before either the trial judge or a Review Board. Section 672.54 of the Criminal Code lists the dispositions available. These range from an absolute discharge to confinement in a hospital. A person found NCR-MD must be granted an absolute discharge if the court or review board is not satisfied that he or she poses a “significant risk”: see Winko v. British Columbia (Forensic Psychiatric Institute) 1999 CanLII 694 (S.C.C.), (1999), 135 C.C.C. (3d) 129 (S.C.C.).

[60] The distinction between non-mental disorder automatism and mental disorder automatism depends on whether the automatistic state is the product of a “mental disorder”. That term is defined in s. 2 of the Criminal Code as a “disease of the mind”. That phrase, which is almost as old as the insanity defence itself, describes a legal and not a medical concept, the purpose of which is normative, not diagnostic: Rabey, per Martin J.A. at pp. 473-74; Parks, per La Forest J. at p. 304. I will use the two phrases interchangeably.

[61] The broader the definition of mental disorder, the narrower the ambit of the “defence” of non-mental disorder automatism. Canadian courts have adopted a very broad definition. In R. v. Cooper reflex, (1980), 51 C.C.C. (2d) 129 (S.C.C.), at p. 144, Dickson J. said:
In summary, one might say that in a legal sense “disease of the mind” embraces any illness, disorder or abnormal condition which impairs the human mind and its functioning, excluding however, self-induced states caused by alcohol or drugs, as well as transitory mental states such as hysteria or concussion.

[62] Professor Brudner has described the effect of this broad definition in these terms:
Clearly the definition of mental disease elaborated by the courts in the aforementioned cases bears little resemblance to any that a psychiatrist might proffer. It is a legal rather than a medical definition – one carefully crafted with a view to a policy of controlling persons thought to be dangerous. This definition ensures, first of all, that anyone accused of a crime who lacked conscious choice because of a dangerous and potentially recurrent disorder or event will not go free but will be subject to continued detention and confinement after acquittal. Obversely, it ensures that only those whose lack of conscious choice was caused by an external and probably non-recurrent event will have the benefit of a defence that leads to an absolute acquittal and a return to society.[4] [Emphasis added.]

[63] The broad definition of the term “mental disorder” has led to the channelling of most automatism claims into the NCR-MD pool. That trend reached its high water mark in Stone. Bastarache J., for the majority, went so far as to take judicial notice “that it will only be in rare cases that automatism is not caused by mental disorder”.[5] Bastarache J. further observed, at para. 199, that trial judges should start with the assumption that the condition constitutes a disease of the mind. Trial judges should then look to the evidence to determine whether that presumption has been rebutted. After Stone, many argue that successful claims of non-mental disorder automatism will be limited to those very rare “one off” cases in which an accused suffers a single incident of automatism, and where
the accused can point to some specific external event that precipitated that event, can demonstrate that the event is unlikely to reoccur, and finally, can show that the event could have produced a disassociative state in an otherwise “normal” person.
(ii) R. v. Parks: Parasomnia and Criminal Responsibility

[64] For over 100 years, leading criminal law jurists and academics accepted without question that parasomnias, like sleepwalking, constituted the prototypical example of non-insane automatism warranting a full acquittal. For example, Stephen J. in R. v. Tolson, [1889] 23 Q.B.D. 168 (C.C.R.), at p. 187, observed:
Can anyone doubt that a man who, though he might be perfectly sane, committed what would otherwise be a crime in a state of somnambulism, would be entitled to be acquitted? And why is this? Simply because he would not know what he was doing.

[65] The words of Stephen J. were repeated and accepted throughout the Commonwealth in the ensuing century: see Bratty, per Viscount Kilmuir at p. 403, per Lord Denning at p. 409; Falconer, per Mason C.J. (in dissent on another point) at p. 43, per Toohey J. at pp. 72-75; Cottle, at p. 1007; Rabey, per Martin J.A. at p. 477, per Dickson J. (in dissent on another issue) at p. 10; Glanville L. Williams, Textbook of Criminal Law, 2d ed. (London: Stevens & Sons, 1983), at p. 665; Peter Ridgway, “Sleepwalking – Insanity or Automatism” (1996) 3 Murdoch U.E.J.L. 4, at para. 4; Irene MacKay, “The Sleepwalker is Not Insane” (1992) 55 Mod. L. Rev. 714, at pp. 715-16.

[66] Parasomnias seemed to withstand the trend towards the treatment of virtually all automatistic behaviour as the product of a disease of the mind. However, in none of the cases in which the courts accepted that parasomnia amounted to non-mental disorder automatism did the accused actually advance parasomnia as an explanation for his or her automatism. The references to parasomnia in the judgments were obiter and were usually by way of explaining why the automatism defence actually advanced by the accused should not result in an acquittal. Put bluntly, in none of these cases was the court faced with the prospect of acquitting a parasomniac who had committed what would otherwise be a serious crime.[6] Furthermore, again because the cases did not actually involve parasomnia, the courts never heard any evidence about the nature of parasomnia.[7]

[67] R. v. Parks put the long held judicial assumption that automatism resulting from parasomnia should lead to an acquittal to a severe test.[8] The accused in Parks claimed that he was in a parasomniac state for a prolonged time period during which he engaged

in a series of activities including leaving his own home, getting into his car, driving across town on a busy highway to the home of his in-laws, entering their home, and attacking them while they were asleep. The accused killed his mother-in-law and seriously injured his father-in-law and then drove to the police station covered in blood. The defence led uncontradicted expert evidence to the effect that the accused’s condition rendered him incapable of controlling his actions and did not constitute a mental disorder. The trial judge did not leave NCR-MD as a possible verdict but instructed the jury that they should acquit if they had a reasonable doubt as to the voluntariness of the accused’s actions. The jury acquitted and the Crown appealed. This court affirmed the acquittal, and on a further appeal to the Supreme Court of Canada that court unanimously affirmed the acquittal.

[68] For present purposes, I will focus on the reasons of La Forest J., delivered for a six-person majority in the Supreme Court of Canada. La Forest J. began by recognizing that the appeal turned on the characterization of the respondent’s automatistic state as either “insane” or “non-insane” automatism. Just as in this case, the accused had led evidence that his actions were involuntary and claimed an acquittal based on that evidence. Again, just as in this case, the Crown denied that the actions were involuntary and alternatively argued that if the actions were involuntary, they were the product of a mental disorder and should result in an insanity verdict and not an acquittal. La Forest J. observed that as the Crown had raised the insanity issue in response to the accused’s contention that his actions were involuntary, the Crown had the onus to prove that the accused’s condition stemmed from a disease of the mind: Parks, at pp. 304, 311. In placing the onus on the Crown, La Forest J. relied on s. 16(4) of the Criminal Code (now s. 16(3)).

[69] Having placed the onus on the Crown, La Forest J. turned to the proper characterization of sleepwalking. He did not simply adopt the cases that had described parasomnia as a self-evident example of non-insane automatism but instead held that parasomnia may or may not be a mental disorder depending on the evidence led in a particular case: Parks, at p. 311.[9]

[70] La Forest J. emphasized, at pp. 304-305, that the phrase “disease of the mind” was a legal and not a medical term. The experts described Parks’ disassociative mental state as the product of sleep and not the product of sleepwalking, an admittedly abnormal condition. According to the experts, sleep was a normal process that could not be characterized as a “disease of the mind”.[10]

[71] La Forest J. did not stop at the experts’ characterization of the accused’s disassociative state. He emphasized that a proper characterization demanded a consideration of factors beyond those relevant to the medical diagnosis, stating at p. 302:
In distinguishing between automatism and insanity the trial judge must consider more than the evidence; there are overarching policy considerations as well. Of course, the evidence in each case will be highly relevant to this policy inquiry.


[72] La Forest J. next observed, at p. 305, that the policy component of the definition of “disease of the mind” focused on the protection of the public. He referred with approval to the observations of Martin J.A. in Rabey, at p. 473:
The legal or policy component relates to (a) the scope of the exemption from criminal responsibility to be afforded by mental disorder or disturbance, and (b) the protection of the public by the control and treatment of persons who have caused serious harms while in a mentally disordered or disturbed state.

[73] La Forest J. and Martin J.A. both acknowledged that public safety considerations may dictate that any abnormal mental state rendering a person incapable of controlling his or her conduct should be characterized as a disease of the mind regardless of the medical characterization or diagnosis. The legal characterization of the condition as a “disease of the mind” places the accused within the s. 16 exemption to criminal liability and forecloses an outright acquittal. Those who are exempted from criminal liability under s. 16 are subject to a post-verdict inquiry into their dangerousness.

[74] La Forest J. identified two approaches that have been used to distinguish non-insane automatism from insane automatism. One approach looked to the cause of the condition and differentiated between causes internal to an accused’s physical or emotional make-up and causes that were entirely external to the accused. The second approach examined the extent to which the accused, because of the nature of the condition, posed a continuing danger to the community. He said at p. 306:
The “continuing danger” theory holds that any condition likely to present a recurring danger to the public should be treated as insanity. The “internal cause” theory suggests that a condition stemming from the psychological or emotional make-up of the accused, rather than some external factor, should lead to a finding of insanity. The two theories share a common concern for recurrence, the latter holding that an internal weakness is more likely to lead to recurrent violence than automatism brought on by some intervening external cause. [Emphasis added.]

[75] La Forest J. indicated that neither approach necessarily provided a definitive answer to the characterization problem presented in any given case. Rather, the two approaches should be treated as analytical tools to be used in answering the ultimate policy inquiry – does the protection of the public dictate that the accused’s disordered mental state should be treated as a disease of the mind?
[76] La Forest J. applied these analytical tools to the evidence. He concluded, at p. 307, that the distinction between internal and external causes was not helpful in characterizing parasomnias. With respect to the ongoing danger, if any, posed by the accused, La Forest J. observed at p. 310:
It is clear from the evidence that there is almost no likelihood of recurrent violent somnambulism. A finding of insanity is therefore less likely, but the absence of a continuing danger does not mean that the respondent must be granted an absolute acquittal.

[77] After determining that neither the causal or ongoing danger analysis provided any clear insight as to the proper characterization of sleepwalking, La Forest J. adverted to additional potentially relevant policy considerations. These considerations proceed from the premise that some automatism claims are feigned and that if automatism claims are too readily accepted by the courts, the administration of justice will be brought into disrepute. These concerns are sometimes described as due administration of justice concerns. I confess to some difficulty in understanding how concerns about the legitimacy of a defence should affect the characterization of an accused’s mental state. In any event, La Forest J. concluded that due administration of justice concerns were irrelevant to the characterization of parasomnia.

[78] La Forest J. ultimately concluded at p. 311:
In the end, there are no compelling policy factors that preclude a finding that the accused’s condition was one of non-insane automatism. I noted earlier that it is for the Crown to prove that somnambulism stems from a disease of the mind; neither the evidence nor policy considerations in this case overcome the Crown’s burden in that regard… [T]he accused should be acquitted.

[79] As is evident, La Forest J. ultimately decided the characterization issue by resort to the burden of proof. On La Forest J.’s analysis, the accused succeeded not because he showed his condition to be one of non-insane automatism but because the Crown failed to prove that the condition should be characterized as insane automatism.

(iii) R. v. Stone, Automatism Reconsidered

[80] I come now to R. v. Stone. The accused stabbed his wife 47 times. At his murder trial, he claimed that he was in a disassociative state brought on by the “psychological blows” inflicted by his wife. According to the appellant, in the hours prior to the homicide, his wife had made many insulting and derisive comments toward him. The accused argued that his actions were involuntary and not the product of a disease of the mind. At trial, the defence sought an acquittal and alternatively a finding of NCR-MD. The jury convicted the accused of manslaughter and the British Columbia Court of Appeal upheld the conviction.

[81] Bastarache J., at p. 419, speaking for a five-person majority, announced that he would “develop a general test applicable to all claims of automatism.” He proceeded to substantially rewrite the case law. Bastarache J. began by determining that the onus of proof should be on the accused to demonstrate on the balance of probabilities that his actions were involuntary. This reversal of the onus of proof on an essential element of the offence is perhaps the most controversial aspect of the ruling in Stone. It is not, however, directly germane to the outcome of this appeal.

[82] The second part of the reasons of Bastarache J. is central to the outcome of this appeal. Bastarache J. held that in those cases where an accused had established a proper evidentiary basis to put voluntariness in issue, the trial judge was required to decide at the end of the evidence whether the claim amounted to one of non-mental disorder automatism or mental disorder automatism. If the claim was properly characterized as mental disorder automatism an acquittal was not available and the trier of fact should be instructed only on the defence of not criminally responsible by reason of mental disorder.

[83] Bastarache J. acknowledged that the distinction between non-mental disorder automatism and mental disorder automatism depended on the definition of the phrase “disease of the mind”. After referring to the position taken by the Canadian Psychiatric Association before a committee of the House of Commons to the effect that all automatism claims should be treated as mental disorder claims, Bastarache J. said at pp. 432-33:
I take judicial notice that it will only be in rare cases that automatism is not caused by mental disorder. Indeed, since the trial judge will have already concluded that there is evidence upon which a properly instructed jury could find that the accused acted involuntarily on a balance of probabilities, there is a serious question as to the existence of an operating mind by the time the disease of the mind is considered. The foregoing lends itself to a rule that trial judges start from the proposition that the condition the accused claims to have suffered from is a disease of the mind. They must then determine whether the evidence in the particular case takes the condition out of the disease of the mind category.

[84] Bastarache J. took the position that anyone who committed what would otherwise be a criminal act and led evidence capable of establishing that his or her actions were involuntary was presumptively suffering from a disease of the mind. This presumption fundamentally changes the legal landscape set out in Parks where, it will be recalled, the non-mental disorder automatism claim succeeded because the Crown could not prove that the accused’s condition constituted a disease of the mind.

[85] Like La Forest J. in Parks, Bastarache J. would not accept that any single approach or criterion could be used to distinguish between non-mental disorder automatism and mental disorder automatism. Bastarache J. described a “holistic approach” that took into consideration the cause of the condition, the continuing danger presented to the public, the due administration of justice policy concerns identified in earlier cases and any other factors made germane to the inquiry by the facts of the particular case. He said at p. 441:
[T]he fundamental question of mixed law and fact which is at the centre of the disease of mind inquiry: whether society requires protection from the accused and, consequently, whether the accused should be subject to evaluation under the regime contained in Part XX.I of the (Criminal) Code. [Emphasis added.]

[86] Bastarache J.’s consideration of the continuing danger factor is germane to this appeal. He observes at p. 439:
In examining the continuing danger factor, trial judges may consider any of the evidence before them in order to assess the likelihood of recurrence of violence. However, two issues will be particularly relevant to the continuing danger factor: the psychiatric history of the accused and the likelihood that the trigger alleged to have caused the automatistic episode will recur. [Emphasis added.]

[87] Bastarache J. explained that a documented history of automatistic-like disassociative states suggested that the condition suffered by the accused was a recurring one. The likelihood of recurrence and more specifically the recurrence of violence during an automatistic state was also enhanced by the fact that on at least one prior occasion, the occasion leading to the criminal charge, the accused had engaged in violent conduct while in an automatistic state. The prior history combined with the prior violent incident indicated that the accused’s condition was “likely to be classified as a disease of the mind”: Stone, at p. 439.

[88] In addressing the significance of the likelihood of the recurrence of the conditions that triggered the automatistic state, Bastarache J. noted that it was difficult to predict the actual recurrence of violence. However, the predictability of the automatistic state in which the violence occurred depended on the likelihood of the recurrence of the triggering events. He said at p. 440:
[A]n assessment of the likelihood that the particular accused will again encounter the trigger alleged to have caused the current automatistic episode, or a similar one of at least equal severity, may assist a judge in assessing the continuing danger factor. The greater the anticipated frequency of the trigger in the accused’s life, the greater the risk posed to the public and, consequently, the more likely it is that the condition alleged by the accused is a disease of the mind.[11] [Emphasis added.]

[89] Bastarache J. then applied his “holistic approach” to the evidence in the case before him. He stressed the absence of any extraordinary external event that could cause a normal person to disassociate. In the absence of any such explanation, the accused’s condition was presumptively a product of a disease of the mind.

[90] Stone alters the approach to the characterization of automatism as non-mental disorder automatism or mental disorder automatism in at least two significant ways. First, after Stone the trial judge must begin from the premise that the automatism is caused by a disease of the mind and look to the evidence to determine whether it convinces him or her that the condition is not a “disease of the mind”. This approach is in direct contrast with Parks where the non-mental disorder automatism claim succeeded because the Crown failed to prove that the condition was caused by a disease of the mind.

[91] Second, although Stone accepts the multi-factored approach to the policy component of the characterization of the automatism set out in Parks, it refocuses the continuing danger aspect of that approach. After Stone, in evaluating the risk of repetition and hence the danger to the public, trial judges must not limit their inquiry only to the risk of further violence while in an automatistic state. Rather, trial judges must examine the risk of the recurrence of the factors or events that triggered the accused’s automatistic state. Commenting on this refinement of the continuing danger inquiry Professor Paciocco observes in “Death by Stone-ing: The Demise of the Defence of Simple Automatism” at p. 281:
This part of the judgment effectively reverses Parks. The triggers for Parks’ somnambulism or sleep-walking included stress, fatigue, insomnia and exercise. There is no point in speaking of the likelihood of such triggers being present in the future. It is a veritable certainty that they will be. It is clear that had Parks been tried using the Stone test, the only defence that would have been left to the jury would be “mental disorder automatism”. [Emphasis added.]

[92] Professor Paciocco’s prediction is largely borne out by the Canadian parasomnia cases that post-date Stone. I am aware of five including this case. In the other four cases, the automatistic states flowing from the parasomnia were held to constitute diseases of the mind: see Canada v. Campbell 2000 CanLII 22824 (ON S.C.), (2000), 35 C.R. (5th) 314 (Ont. S.C); R. v. Balenko, [2000] Q.J. No. 717 (C.Q. (Crim. Div.)); R. v. Romas (2002), 6 M.V.R. (5th) 101 (B.C. Prov. Ct.); and R. v. Churchyard, an unreported decision of Smith J. released November 19, 2003 (Ont. S.C.).

[93] The majority position in Stone signals a strong preference for a finding of NCR-MD in cases where an accused establishes that he or she was in a disassociative state and acted involuntarily. Social defence concerns, inevitably present in such cases, must to a large degree drive the analysis in automatism cases after Stone.

[94] The strong preference for an NCR-MD verdict expressed in Stone is explained in part by the very different treatment accorded those found NCR-MD compared to the historical treatment provided to those found not guilty by reason of insanity, as was the case at the time of the trial in Parks. Prior to 1991, persons found not guilty by reason of insanity were detained indefinitely at the pleasure of the Lieutenant Governor in Council. The provisions of Part XX.I of the Criminal Code not only disposed of the insanity nomenclature but completely changed the post-verdict treatment of those found NCR-MD: S.C. 1991, c. 43.

[95] I do not think it is coincidental that Stone and Winko, the leading case on the interpretation of Part XX.I of the Criminal Code, were heard by the same nine judges about ten days apart, and decided about three weeks apart several months later. It is hard to resist the inference that Stone was written having in mind what the court would say three weeks later when it released its decision in Winko.

[96] In Winko, McLachlin J. explained the operation of Part XX.I this way at para. 43:
In summary, the purpose of Part XX.I is to replace the common law regime for the treatment of those who offend while mentally ill with a new approach emphasizing individualized assessment and the provision of opportunities for appropriate treatment. Under Part XX.I, the NCR accused is neither convicted nor acquitted. Instead, he or she is found not criminally responsible by reason of illness at the time of the offence. This is not a finding of dangerousness. It is rather a finding that triggers a balanced assessment of the offender’s possible dangerousness and of what treatment-associated measures are required to offset it. Throughout the process the offender is to be treated with dignity and accorded the maximum liberty compatible with Part XX.I’s goals of public protection and fairness to the NCR accused. [Emphasis added.]

[97] Winko provides a detailed examination of s. 672.54, the provision governing the dispositions available with respect to persons found NCR-MD. As interpreted in Winko, s. 672.54 requires the absolute discharge of anyone found NCR-MD unless the court or the Review Board determines that the individual poses “a significant threat to the public”. McLachlin J. said at para. 52:
This interpretation of s. 672.54 eliminates any need for the NCR accused to prove lack of dangerousness and relieves him or her of any legal or evidentiary burden. If the evidence does not support the conclusion that the NCR accused is a significant risk, the NCR accused need do nothing; the only possible order is an absolute discharge.

[98] The risk determination required by s. 672.54 cannot not be based on speculation or assumptions about how persons with mental disorders behave. There must be evidence establishing the significant risk. That risk must be a real risk of criminal conduct involving physical or psychological harm to individuals in the community. A risk of trivial harm or miniscule risk of significant harm will not suffice to deprive the individual of his or her liberty: Winko, at para. 57.

[99] As explained in Winko, there are also significant procedural safeguards built into Part XX.I of the Criminal Code. A person found NCR-MD must receive a timely disposition hearing before either the court or the Review Board. He or she has full access to the bail provisions pending that hearing. The disposition hearing is not adversarial but provides for a full and wide-ranging inquiry into all factors relevant to the appropriate disposition. The disposition hearing allows for the input of medical professionals who will have had a chance to assess the person found NCR-MD and to develop opinions based on an up-to-date assessment of that person’s condition. Where, as in most cases, the disposition hearing is before the Review Board, the panel will include a psychiatrist.

[100] A combined reading of Stone and Winko yields a comprehensive response to automatism claims. At the pre-verdict stage, social defence concerns dominate. Those concerns focus on the risk posed by the potential recurrence of the conduct in issue. Where that risk exists, the risk combined with the occurrence of the conduct that led to the criminal proceedings will almost always justify further inquiry into the accused’s dangerousness so as to properly protect the public.

[101] In the post-verdict stage, however, the emphasis shifts to an individualized assessment of the actual dangerousness of the person found NCR-MD. Where that personalized assessment does not demonstrate the requisite significant risk, the person found NCR-MD must receive an absolute discharge. Even where a significant risk exists, the disposition order must be tailored to the specific circumstances of the individual and must, to the extent possible, minimize the interference with that individual’s liberty.

[102] Re Romas, [2002] B.C.R.B.D. No. 66 (British Columbia Review Board), provides an example of the integration of the Stone approach to automatism and the Winko approach to Part XX.I of the Criminal Code. At trial (R. v. Romas), the accused was charged with an assault-related offence and claimed to have acted as a result of Confusional Sleep Arousal, a sleep disorder. Applying Stone, the accused was found NCR-MD. The trial judge referred the matter to the British Columbia Review Board for disposition. About six weeks later the Board, by a majority vote, ordered the accused absolutely discharged. In so ordering, the Board referred to several factors, all of which would seem applicable to the respondent assuming his current condition is consistent with that described at trial.
(iv) The Trial Judge’s Application of the Principles in R. v. Stone

[103] I am satisfied that the trial judge failed to appreciate the significance of the hereditary nature of the respondent’s condition, failed to give effect to the respondent’s well established history of sexsomnia, and failed to appreciate the significance of the strong likelihood of the recurrence of the events that triggered his sexsomnia. The trial judge also failed to appreciate that Dr. Shapiro’s medical opinion that parasomnia did not constitute a mental disorder was largely irrelevant to the determination of whether, for policy reasons, the condition should be classified legally as a disease of the mind. These errors led to a failure to apply the proper legal standard when characterizing the respondent’s automatism.

[104] When addressing the cause of the respondent’s condition, the trial judge acknowledged that causes internal to the accused, be they physical or mental, were generally treated as indicative of a disease of the mind, while causes external to the accused were generally indicative of a non-mental disorder automatism. The trial judge then said at para. 46:
The jurisprudence recognizes that somnambulism, as in the Parks case, is not suitable to this kind of analysis. Somnambulism is not a disease of the mind.

[105] The trial judge was wrong to categorically indicate that somnambulism is not a disease of the mind. As Parks made clear, it may or may not be depending on the evidence. More significantly to the outcome of this case, I think the trial judge erred in dismissing the potential significance of the causes of the respondent’s condition by referring to jurisprudence that had dismissed the internal/external cause distinction as being of no assistance in parasomnia cases. The trial judge should have considered whether, on the evidence, the causes of the parasomnia as explained by Dr. Shapiro offered any insight into the risk of recurrence and hence the potential danger to the public.

[106] While the cause of parasomnia may not fit within the “external/internal” causal dichotomy described in the case law, Dr. Shapiro’s evidence establishes that the predisposition for parasomnia, found in some three per cent of the adult population, is hereditary. A genetic predisposition is the epitome of an internal cause. Although that disposition does not cause the particular automatistic event, it does predispose the individual to that condition thereby increasing the risk of recurrence. The trial judge erred in discounting the significance of this internal cause of the respondent’s condition. He did so based on a misapprehension of the “sleepwalking” case law and a failure to consider evidence relevant to the causal inquiry.

[107] The trial judge’s treatment of the “continuing danger” component of the meaning of “disease of the mind” reveals two errors. First, he found that prior sexsomnia incidents did not assist in assessing the danger to the public because those incidents did not involve criminal conduct. While technically the conduct did involve non-consensual sexual activity, I certainly accept that it was not conduct that should attract a criminal charge.

[108] Even though the incidents involving former girlfriends did not lead to criminal charges, they were important in assessing the risk posed by the respondent’s conduct. These events correspond to the psychiatric history of prior automatistic episodes stressed in Stone. The respondent’s prior sexsomnia episodes demonstrate that his conduct toward L.O. was not an isolated, “one-off” incident. There were episodes both prior to and subsequent to the attack on L.O. during which the respondent engaged in sexual activity at a time when he was unaware of what he was doing and unable to control himself. In that condition the respondent could not distinguish between a consenting partner and a non-consenting victim. The identity of the subject of the sexual activity was a matter of chance, not a matter of choice.

[109] I accept that it is perhaps more probable that the respondent would be asleep with or near someone with whom he had a relationship involving consensual sexual relations. It cannot, however, be unexpected that the respondent will find himself asleep in the vicinity of persons with whom he has no relationship. One thinks of an airport or train station waiting room or the sleeping arrangements at a friend’s cottage. There is also nothing in the evidence to suggest that the respondent could not move from room to room while in a parasomniac state and engage in non-consensual sexual activity with someone who happens to be in same building but in a different room.

[110] With respect to the trial judge, his rejection of the prior sexsomnia incidents as having no value in considering whether the respondent posed a risk runs contrary to the fundamental nature of the defence advanced by the respondent. It was central to his position that he had no awareness of what he was doing and no control over his actions. It must follow that he was incapable of distinguishing between a consensual or non-consensual situation while in a parasomniac state. The trial judge should not have relied on a distinction that the accused was incapable of making to diminish the risk posed by the respondent.

[111] The second error made by the trial judge in examining the “continuing danger” factor arises from his failure to consider the likelihood that the respondent would encounter the events or circumstances that triggered his parasomnia. Stone, at p. 440, explicitly directs trial judges to look to the likelihood of the recurrence of the triggering events and not just the likelihood of the recurrence of acts of violence while in an automatistic state.

[112] The triggers in this case included alcohol, fatigue and stress. These are common place in most people’s lives, particularly in the life of a busy, socially active, young man. Had the trial judge addressed the likelihood of the recurrence of the triggering events he would no doubt have found that it was virtually inevitable that some combination of these events would recur and recur with some frequency. On the analysis in Stone, the likelihood of recurrence of these triggering events offers significant support for a finding that the respondent’s parasomnia should be characterized as a disease of the mind.

[113] The trial judge also placed considerable emphasis on the firmness of Dr. Shapiro’s evidence that the respondent’s parasomnia did not constitute a disease of the mind. Dr. Shapiro’s opinion that parasomnia did not constitute a disease of the mind, as opposed to his explanation of parasomnia and his opinion that the respondent was in a parasomniac state, had little or no evidentiary value. Nor was the force with which he held that opinion of any significance. As Martin J.A. observed in Rabey, at p. 474:
Indeed, in strictness, a medical witness is not entitled to state that a particular condition is or is not a disease of the mind since that is a legal question. In practice, however, it is often convenient and permissible, in the discretion of the Judge, for a medical witness to testify in those terms.

[114] The trial judge should not have been influenced by Dr. Shapiro’s opinion as to the proper categorization of the respondent’s medical condition when deciding whether as a matter of law and for policy reasons that condition should or should not be characterized as a disease of the mind.

[115] Mr. Addario for the respondent ably, although ultimately unsuccessfully, defended the reasoning of the trial judge. Near the end of his oral submissions he opened a second front in support of his position that the appeal must be dismissed. Mr. Addario submitted that even if a strict application of the analysis set out in Stone might suggest an NCR-MD verdict, it would be grossly unfair and unjust to the respondent to label him NCR-MD. Mr. Addario submitted that as wrongheaded as it no doubt is, there is a very strong negative stereotype of persons found to be NCR-MD. Mr Addario submits that the reality of the respondent’s medical condition and his life in general could not be further from that stereotype. He contends that the respondent’s parasomnia would not be regarded by any reasonable member of the community as the kind of mental disorder associated with a finding of NCR-MD. Mr Addario forcefully argues that, just as the criminal justice system must be concerned with illegitimate automatism claims, it must be concerned that verdicts in automatism cases bear some resemblance to the community’s concept of those who should or should not be found NCR-MD.

[116] There is considerable force in Mr. Addario’s submissions. Clearly no reasonable member of the community would, or should, regard the respondent as mentally ill. Indeed, the inappropriateness of labelling persons with parasomnia as mentally ill was acknowledged by Chief Justice Lane in Burgess, at p. 776.

[117] I agree with Mr. Addario’s contention that the result in criminal proceedings should, at least on a general level, reflect the community’s perception of the appropriate result. Proper labelling is important to the maintaining of the integrity of the criminal justice process. It would be preferable if the Criminal Code allowed the court, where it was deemed necessary, to specifically identify in its verdict the condition that caused the involuntary actions, e.g. not criminally responsible on account of parasomnia: see Ebrahim, et al., at p. 612. Unfortunately the Criminal Code does not provide for that flexibility. Arguably, neither of the available verdicts, not guilty or NCR-MD, would be regarded by the community as particularly apt labels for the respondent.

[118] The second part of Mr. Addario’s submission would have the court accept the negative stereotyping of those found NCR-MD and decline to impose that verdict on the respondent because he does not fit that stereotype. No one should deny the existence of this negative stereotype and the harm it can do to those found NCR-MD. To give effect to Mr. Addario’s submission would, however, promote this negative stereotype. Were the court to decline to find the respondent NCR-MD because he does not fit the negative stereotype of persons so found, the court could be taken as accepting that those who are found NCR-MD do fit that stereotype.

[119] The new mental disorder regime introduced into the Criminal Code in 1991 is intended to overcome the improper stereotyping of persons found NCR-MD and to provide for individualized assessment and treatment of those individuals: see Winko, at paras. 35-40. The courts can best play a role in the important task of overcoming the negative stereotypes associated with mental illness, not by shaping their verdicts to conform to those stereotypes, but by emphasizing both the basis for a finding of NCR-MD and by explaining what the verdict means. An NCR-MD verdict signals that an accused cannot be held responsible for what would otherwise be his or her criminal act. At the same time, it rejects any suggestion that the accused represents an automatic danger to the public. Instead, the NCR-MD verdict triggers an individualized, careful, current assessment of the accused’s condition leading to a disposition tailored to the individual accused.

IV. THE APPROPRIATE ORDER

[120] The Crown has established an error in law that vitiates the acquittal. On a proper application of the principles taken from Stone, to the facts as found by the trial judge, a NCR-MD verdict is the only verdict available in law. By analogy to Crown appeals from acquittals by a judge alone where the Crown seeks a conviction on appeal, an order setting aside the acquittal and substituting an NCR-MD verdict would seem to be the correct order: see R. v. Cassidy 1989 CanLII 25 (S.C.C.), (1989), 50 C.C.C. (3d) 193 (S.C.C.), at p. 200. Crown counsel in her original factum and in oral submissions asked the court to allow the appeal, set aside the acquittal and substitute an NCR-MD verdict. The respondent did not suggest that the court did not have jurisdiction to substitute an NCR-MD verdict should it allow the appeal and set aside the acquittal.

[121] While this appeal was under reserve, the panel asked counsel for further written submissions addressing the scope of the court’s remedial powers should it allow the appeal and set aside the acquittal. In her written submissions, Ms. Crosbie took the position that the court cannot substitute an NCR-MD verdict, and under the terms of s. 686(4) can only order a new trial. She further submitted that s. 686(8) does not allow this court to limit the new trial to the determination of whether the respondent’s automatism constituted non-mental disorder automatism resulting in an acquittal, or mental disorder automatism resulting in an NCR-MD verdict. In their written reply, counsel for the respondent agreed that this court cannot substitute an NCR-MD verdict. They contend, however, that the court can and should make an order under s. 686(8) limiting the scope of the new trial. Counsel argued that the Crown should not be permitted to re-litigate the voluntariness of the respondent’s conduct. The new trial should be directed only at determining whether that involuntary conduct should result in an acquittal or an NCR-MD verdict.

[122] Appeals are creatures of statutes. The powers of the Court of Appeal on appeals in indictable matters are found in Part XXI of the Criminal Code: see R. v. Thomas 1998 CanLII 774 (S.C.C.), (1998), 130 C.C.C. (3d) 225 (S.C.C.), at para. 14; R. v. Wells 2004 ABCA 371 (CanLII), (2004), 193 C.C.C. (3d) 43 (Alta. C.A.), at para. 12. Section 676(1)(a) creates the Crown’s right of appeal from an acquittal or an NCR-MD verdict.
The Attorney General or counsel instructed by him for the purpose may appeal to the court of appeal
(a) against a judgment or verdict of acquittal or a verdict of not criminally responsible on account of mental disorder of a trial court in proceedings by indictment on any ground of appeal that involves a question of law alone

[123] Section 686(4) sets out the Court of Appeal’s remedial powers on appeals from an acquittal or a finding of NCR-MD:
If an appeal is from an acquittal or verdict that the appellant or respondent was unfit to stand trial or not criminally responsible on account of mental disorder, the court of appeal may

(a) dismiss the appeal; or
(b) allow the appeal, set aside the verdict and
(i) order a new trial, or
(ii) except where the verdict is that of a court composed of a judge and jury, enter a verdict of guilty with respect to the offence of which, in its opinion, the accused should have been found guilty but for the error in law, and pass a sentence that is warranted in law, or remit the matter to the trial court and direct the trial court to impose a sentence that is warranted in law.

[124] Section 686(4) specifically provides for two remedies if a Crown appeal from acquittal is successful. The court may order a new trial, or if the original verdict was not returned by a jury, substitute a conviction. Section 686(4) says nothing about the court’s power to substitute an NCR-MD verdict for an acquittal. Nor is there anything in the language of s. 686(4) to suggest that the power to substitute an NCR-MD verdict is implicit in the language of that section.

[125] The absence of any reference in s. 686(4) to the power to substitute an NCR-MD verdict for an acquittal takes on added significance when one examines the powers of the Court of Appeal on appeals from conviction. Section 686(2)(b) provides that the court can set aside a conviction and order a new trial. A different section, s. 686(1)(d), expressly allows a court of appeal to substitute an NCR-MD verdict for a conviction where that conviction is set aside on a conviction appeal. In R. v. Warsing 1998 CanLII 775 (S.C.C.), (1998), 130 C.C.C. (3d) 259 (S.C.C.), Major J., for the majority, read the two remedial sections as providing alternative remedies on an appeal from conviction where an appellant’s sanity was challenged for the first time on appeal. After examining the two provisions, he said at para. 68:
Without s. 686(1)(d) a court of appeal would not be able to substitute a verdict of NCRMD as s. 686(1)(a) and s. 686(2) [the new trial provisions] alone do not provide for this remedy.

[126] There is no equivalent to s. 686(1)(d) applicable to appeals from acquittals. On the reasoning in Warsing, this court can only order a new trial and cannot substitute an NCR-MD verdict on a successful Crown appeal from an acquittal.

[127] Section 686(4) is far from a masterpiece of legislative draftsmanship. As pointed out in R. v. Yelle 2006 ABCA 276 (CanLII), (2006), 213 C.C.C. (3d) 20 (Alt. C.A.), the section provides no express remedies on a successful Crown appeal from a trial judgment that had ordered a stay of proceedings. In Yelle, Martin J.A., at paras. 34-38, expressed the opinion that the absence of any remedial provision reflected a clear legislative oversight which if left uncorrected would cause significant harm to the administration of justice. Martin J.A. read the statutory power to order a new trial found in s. 686(4) as including the power to make an order directing the continuation of the initial trial where the Crown appeal was from an order staying that trial.

[128] The reasoning in Yelle is compelling, however, I do not think I can apply that reasoning to this case. Parliament has clearly turned its mind to the circumstances in which the Court of Appeal may substitute an NCR-MD verdict. It has chosen to extend that power on conviction appeals but not on appeals from acquittals. An appellate court’s remedial powers on Crown appeals have traditionally been much more circumscribed than appellate powers on conviction appeals. I doubt that the absence of a power to substitute an NCR-MD verdict on a Crown appeal from an acquittal in a non-jury case reflects a considered policy decision. However, I cannot say that it clearly demonstrates a legislative oversight justifying a judicial rewrite of the provision.

[129] Section 686(4) by its clear language provides for only two remedies on appeals from acquittals entered by judges sitting without juries. This court may order a new trial, or enter a conviction. Obviously this is not a case for a conviction. A new trial is the only possible remedy.
[130] I come now to whether this court can make an order limiting the issues at the new trial. While counsel disagree as to the existence of the power to limit the new trial, they do agree that if the power exists it must be found in s. 686(8):
Where a court of appeal exercises any of the powers conferred by subsection (2), (4), (6) or (7), it may make any order, in addition, that justice requires.

[131] Section 686(8) creates a broad discretionary power to make a wide variety of orders ancillary to the primary order made under the remedial provisions in s. 686: R. v. P. (D.W.) 1989 CanLII 71 (S.C.C.), (1987), 49 C.C.C. (3d) 417 (S.C.C.), at pp. 419-20; Thomas, at para. 19. There are, however, limits on the reach of s. 686(8). It does not create a free standing remedial power but provides a supplementary power that may be used, where justice requires, to augment the remedial powers found elsewhere in s. 686. As a supplementary power, an order made under s. 686(8) cannot be at variance with the primary order made in any given case. Nor can an order made under s. 686(8) be inconsistent with the remedial powers granted under the other provisions in s. 686. For example, s. 686(8) could not be used to substitute a conviction on a Crown appeal from an acquittal entered by a jury: see R. v. Pearson 1998 CanLII 776 (S.C.C.), (1998), 130 C.C.C. (3d) 293 (S.C.C.), at para. 14; Thomas, at paras. 17, 21-22.

[132] This court has previously held that the remedial power in s. 686(8) is broad enough to permit, in unusual circumstances, an order limiting the scope of a new trial. In Reference Re Regina v. Gorecki (No. 2) (1976), 32 C.C.C. (2d) 135, the court received fresh evidence directed at the sanity of the accused at the time of the homicide. The court concluded that a new trial should be directed and that the new trial should be limited to the insanity issue. In R. v. Wade 1992 CanLII 748 (ON C.A.), (1994), 89 C.C.C. (3d) 39, reversed on other grounds 1995 CanLII 100 (S.C.C.), (1995), 98 C.C.C. (3d) 97 (S.C.C.), this court directed a new trial limited to whether the accused was guilty of murder or manslaughter. In Wade the accused claimed that he was in an automatistic state at the time of the homicides as a result of certain sleep disorders. The jury rejected that claim and convicted the accused. The court held that there was non-direction on the included offence of manslaughter. It further held, however, that the issue of the voluntariness of the accused’s conduct had been fully and fairly litigated. Consequently the court, relying on s. 686(8), limited the new trial to the question of whether the accused was guilty of murder or manslaughter and foreclosed any further litigation on the voluntariness issue.

[133] The two decisions referred to above offer direct support for the respondent’s contention that the scope of a new trial ordered in this case can be limited to the issue of the nature of the respondent’s automatism. The Crown argues, however, that Gorecki and Wade have been implicitly but clearly overruled by the Supreme Court of Canada in Thomas and Warsing. In Thomas, the majority of the court quashed the order of the British Columbia Court of Appeal limiting the new trial to the question of whether the accused was guilty of murder or manslaughter. The majority directed a full new trial. In Warsing, the majority reversed the decision of the British Columbia Court of Appeal limiting the new trial to the issue of the accused’s sanity at the time of the homicide. The court ordered a full new trial.

[134] A reading of the majority judgments in Thomas and Warsing, in combination with the dissents, compels me to agree with the Crown’s submission that neither Gorecki nor Wade reflect the present state of the law as it relates to s. 686(8). It does not follow from that observation, however, that this court has no jurisdiction to limit the scope of the new trial in the circumstances of this appeal.

[135] Both Thomas and Warsing were cases in which the orders made by the appellate court limiting the new trial effectively eliminated the possibility of an acquittal and restricted, if not foreclosed, certain potential defences. Other authorities relied on by the Crown are also cases in which the proposed limits placed on the new trial would have adversely affected the rights of the accused: see R. v. MacKay 2005 SCC 79 (CanLII), (2005), 203 C.C.C. (3d) 289 (S.C.C.); Wells, at para. 19.

[136] I accept counsel for the respondent’s submission that the considerations that led the Supreme Court of Canada in Warsing and Thomas to conclude that there was no jurisdiction in s. 686(6) to limit the scope of a new trial do not apply where the restriction on the scope of the new trial does not disadvantage the accused. In Warsing, at paras. 72-75, Major J. explained that limiting the new trial to the issue of whether the accused was NCR-MD was offensive to s. 686(6) because it would deny the accused the presumption of innocence and the constitutional right to put forward whatever defence the accused deemed appropriate. In the present case, neither the presumption of innocence nor the accused’s right to control his own defence would be adversely affected by an order limiting the respondent’s new trial to a determination of whether his automatism should lead to an acquittal or an NCR-MD verdict. To the extent that the presumption of innocence remains alive in automatism cases after Stone, that presumption is not affected by an order limiting the scope of the new trial. That order would eliminate only the possibility of a conviction (although, for reasons set out above, I think the evidence as it presently stands effectively eliminates the possibility of an outright acquittal).

[137] The accused supports the order limiting the scope of the new trial. He does not seek to re-litigate the voluntariness issue. In those circumstances, it cannot be said that a limit on the scope of the new trial would interfere with the accused’s conduct of his own defence. To the contrary, the proposed order would reflect the manner in which the respondent wishes to conduct his defence. The position of the respondent with respect to the proposed limitation on the new trial distinguishes this case from Wade, where the accused very much wanted to re-litigate the voluntariness issue that he had lost at the initial trial.

[138] The submissions made on behalf of the respondent have convinced me that s. 686(8) does authorize an order ancillary to an order for a new trial made under s. 686(4) limiting the scope of the new trial. That ancillary order may be made only where it does not interfere with any of the accused’s rights and is otherwise consistent with the demands of justice in the circumstances: see R. v. Druken 2002 NFCA 23 (CanLII), (2002), 164 C.C.C. (3d) 115 (Nfld. C.A.), at pp. 139-40. I would think that such orders would be rare.

[139] Apart from the jurisdictional argument, the Crown does not advance any submissions to suggest that an order limiting the scope of the respondent’s new trial would be unfair to the interests represented by the Crown, cause practical problems at the new trial or otherwise interfere with the due administration of justice. The respondent, however, makes a convincing case that it would be unfair to compel him to re-litigate the voluntariness issue. He successfully established on the balance of probabilities that his conduct was involuntary. The trial judge fully and fairly reviewed the evidence before making that finding. His review reveals no error. Forcing a re-litigation of the voluntariness issue over the respondent’s objection many years after the relevant events occurred would be unfair to the respondent and unlikely to produce either substantive or procedural justice.

[140] I would make an order under s. 686(6) limiting the new trial to whether on a proper application of the principles in Stone the respondent’s automatism should result in an acquittal or an NCR-MD verdict. I recognize that in limiting the new trial in the manner I have, there may be little independent judgment for the trial judge to exercise at the new trial. If the evidence is substantially the same as it was at the first trial, an NCR-MD verdict is the only reasonable verdict available from the state of the current law. I would expect the trial judge to reach that verdict, assuming the evidence is substantially the same. The evidence may, of course, be different.

[141] There is some discussion in counsel’s written submissions about the possibility of remitting the matter to the original trial judge for the re-trial. Section 686(5)(b) may apply, in which case the re-trial could be before the original trial judge only if the court so ordered. I do not propose to make that order at this time. I do not, however, suggest that there is necessarily any impediment to the original trial judge conducting the re-trial. If counsel for the parties and the original trial judge are inclined to the view that it would be in the best interests of justice for him to conduct the re-trial I would be prepared to make the order necessary to remove any statutory impediment to him conducting the re-trial.

V. CONCLUSION

[142] I would allow the appeal, set aside the acquittal and order a new trial. Pursuant to s. 686(8), I would limit the scope of the new trial to a determination of whether the respondent’s automatism should result in a verdict of not guilty or an NCR-MD verdict.