R. v. Kruk, 2024 SCC 7

SUPREME COURT OF CANADA
Citation: R. v. Kruk, 2024 SCC 7			Appeals Heard: May 18, 2023 Judgment Rendered: March 8, 2024 Dockets: 40095, 40447
Between:   His Majesty The King Appellant   and   Christopher James Kruk Respondent   - and -   Independent Criminal Defence Advocacy Society, Criminal Lawyers’ Association (Ontario) and Trial Lawyers Association of British Columbia Interveners   And Between:   His Majesty The King Appellant   and   Edwin Tsang Respondent   - and -   Attorney General of Alberta, Independent Criminal Defence Advocacy Society, Association québécoise des avocats et avocates de la défense, West Coast Legal Education and Action Fund Association, Women’s Legal Education and Action Fund Inc. and Trial Lawyers Association of British Columbia Interveners       Coram: Wagner C.J. and Côté, Rowe, Martin, Kasirer, Jamal and O’Bonsawin JJ.
Reasons for Judgment: (paras. 1 to 127)	Martin J. (Wagner C.J. and Côté, Kasirer, Jamal and O’Bonsawin JJ.)
Concurring Reasons: (paras. 128 to 249)	Rowe J.

 

 

Note: This document is subject to editorial revision before its reproduction in final form in the Canada Supreme Court Reports.

 

 

His Majesty The King                                                                                    Appellant

v.

Christopher James Kruk                                                                            Respondent

and

Independent Criminal Defence Advocacy Society,

Criminal Lawyers’ Association (Ontario) and

Trial Lawyers Association of British Columbia                                       Interveners

‑ and ‑

His Majesty The King                                                                                    Appellant

v.

Edwin Tsang                                                                                                Respondent

and

Attorney General of Alberta,

Independent Criminal Defence Advocacy Society,

Association québécoise des avocats et avocates de la défense,

West Coast Legal Education and Action Fund Association,

Women’s Legal Education and Action Fund Inc. and

Trial Lawyers Association of British Columbia                                       Interveners

Indexed as: R. v. Kruk

2024 SCC 7

File Nos.: 40095, 40447.

2023: May 18; 2024: March 8.

Present: Wagner C.J. and Côté, Rowe, Martin, Kasirer, Jamal and O’Bonsawin JJ.

on appeal from the court of appeal for british columbia

                    Criminal law — Appeals — Standard for appellate intervention — Credibility and reliability assessment — Common-sense assumptions — Accused both convicted of sexual assault at trial — Court of Appeal finding that trial judges’ credibility and reliability assessments were based on common-sense assumptions not grounded in evidence — Court of Appeal overturning convictions on basis that trial judges erred in law by failing to abide by rule against ungrounded common-sense assumptions — Whether error of law based on rule against ungrounded common-sense assumptions should be recognized.

                    K and T were convicted of sexual assault in separate and unrelated matters. In both cases, the Court of Appeal overturned the convictions on the basis of alleged errors of law in the trial judges’ credibility and reliability assessments. Using the rule against ungrounded common‑sense assumptions, which originated in a series of appellate cases, the Court of Appeal found that the trial judges erred in law by making assumptions about human behaviour not grounded in the evidence. In K’s appeal, the court held that the trial judge’s conclusion that it was unlikely that a woman would be mistaken about the feeling of penile‑vaginal penetration relied on speculative reasoning and was not the proper subject of judicial notice. In T’s appeal, the court held that the trial judge had made three assumptions about human behaviour that had impacted her assessment of the evidence: (1) a person would not ask to be spanked while engaging in sexual foreplay out of the blue; (2) a controlling person would not refrain from engaging in vaginal intercourse because they could not find a condom; and (3) a person would not abruptly and unceremoniously drive away from the person with whom they had engaged in consensual sex. The court found that these generalizations were not based in the evidence and engaged in speculative reasoning, and that these errors were material. New trials were ordered for K and T.

                    Held: The appeals should be allowed and the convictions restored.

                    Per Wagner C.J. and Côté, Martin, Kasirer, Jamal and O’Bonsawin JJ.: The rule against ungrounded common‑sense assumptions should not be recognized as giving rise to an error of law. Such an error of law would represent a radical departure from how appellate courts have typically approached credibility and reliability assessments, especially in the context of sexual assault. The proposed rule is in no way analogous to the body of law protecting sexual assault complainants from myths and stereotypes, nor can it be justified as necessary to ensure fairness to the accused. It also treats any and all factual assumptions drawn in the course of testimonial assessments as errors of law and thereby represents an unjustified departure from well‑established principles governing testimonial assessment and appellate standards of review. The faulty use of common‑sense assumptions in criminal trials should continue to be controlled by existing standards of review and rules of evidence. In some cases, a trial judge’s use of common sense will be vulnerable to appellate review because it discloses recognized errors of law. Otherwise, like with other factual findings, credibility and reliability assessments — and any reliance on the common‑sense assumptions inherent within them — will be reviewable only for palpable and overriding error. In the instant cases, assessing the trial judges’ credibility and reliability findings using the proper standard of palpable and overriding error, no such errors were made.

                    First, the proposed rule against ungrounded common‑sense assumptions is not a logical extension of the prohibition against myths and stereotypes about sexual assault complainants. It reflects a misunderstanding of the distinct body of law associated with myths and stereotypes in sexual assault cases, which has a unique history and a specific remedial purpose: to remove discriminatory legal rules that contributed to the view that women, as a group, were less worthy of belief and did not deserve legal protection against sexual violence. Several myths and stereotypes have been jurisprudentially condemned as errors of law and significant legislative changes were made with a view to protecting the rights of women and children given their particular vulnerability to sexual violence. This history puts into perspective the distinct reasons why relying on myths and stereotypes to discredit sexual assault complainants amounts to an error of law, as opposed to being an ordinary factual finding reviewable for palpable and overriding error. Conversely, the proposed rule does not relate to specific, identified, erroneous generalizations about a specific category of witness, nor does it protect elements of an offence from taking on a distorted meaning. It instead lumps together the sorts of pernicious, discriminatory stereotypes that both the courts and Parliament have worked to condemn and correct with more benign generalizations that, while they may be factually wrong, have nothing to do with inequality of treatment.

                    The rule would also drastically expand the scope of permissible questioning into a complainant’s sexual history, effectively requiring both parties to apply to adduce other sexual activity evidence that may not otherwise be relevant or permitted, opening a back door to prohibited twin‑myth reasoning. Although framed in terms of ensuring equal treatment for the accused, this approach in fact risks resurrecting the very prejudice against sexual assault complainants that the law on myths and stereotypes was designed to eliminate. Recognizing an identical rule mirroring the treatment of myths and stereotypes between complainants and accused is not necessary and would be misguided. The accused’s rights remain safeguarded by crucial legal protections explicitly designed to ensure fairness to the accused that find their source in their own robust body of law flowing from principles such as the presumption of innocence, the right to silence, and reasonable doubt. Such protections ensure fairness to the accused and must guide trial judges in assessing testimony.

                    Furthermore, the proposed rule is counterproductive to proper testimonial assessment and incompatible with the often inextricable role common‑sense assumptions play in credibility and reliability assessments. By prohibiting ungrounded common‑sense assumptions, the rule interferes with the necessary recourse to common sense as a part of testimonial analysis. It is effectively impossible to draw a clear boundary between using human experience to interpret evidence or draw inferences (which is permissible under the rule) and introducing new considerations into the evidence (which is not). The rule invites appellate courts to substitute their opinions about what generalizations are appropriate or instructive for those of trial judges, improperly transforming their strong opposition to a trial judge’s factual inferences into supposed legal errors, thus creating uncertainty and unfairness on appeal.

                    The rule also runs contrary to established standards of review and would unduly increase the scope of appellate intervention into the credibility and reliability assessments of trial judges. These assessments can be the most important and difficult judicial determinations in a criminal trial, especially in sexual assault cases, which often involve acts that allegedly occurred in private and hinge on the contradictory testimony of two witnesses. Although credibility and reliability findings may be overturned on correctness if errors of law are disclosed, in most cases it is preferable to review them using the nuanced and holistic standard of palpable and overriding error — which defers to the conclusions of trial judges who have had direct exposure to the witnesses themselves and have expertise in assessing and weighing the facts. The reasons for the deference accorded to a trial judge’s factual and credibility findings include: (1) limiting the cost, number, and length of appeals; (2) promoting the autonomy and integrity of trial proceedings; and (3) recognizing the expertise and advantageous position of the trial judge. Appellate courts are comparatively ill‑suited to credibility and reliability assessment, being restricted to reviewing written transcripts of testimony and often focusing narrowly on particular issues as opposed to seeing the case and the evidence as a whole. Invoking the proposed rule, appellate courts have been invited to parse trial reasons, attack generic statements made in the course of credibility assessments, and frame any credibility findings based on human behaviour as impermissible stereotype or common‑sense assumptions untethered to evidence. The jurisprudence in this area is variable, even volatile, and evinces the need for a more consistent approach to appellate review.

                    In view of the rejection of the proposed rule, appellate courts should continue to rely on the existing and well‑established law on assessing a trial judge’s credibility or reliability assessments. First, where an appellant alleges that a trial judge erroneously relied on a common‑sense assumption in their testimonial assessment, the reviewing court should consider whether what is being impugned is, in fact, an assumption. What might appear to be an assumption may actually be a judge’s particular finding about the witness based on the evidence. Second, if the trial judge did rely on an assumption that is beyond the bounds of what common sense and the judicial function support, the reviewing court should identify the appropriate standard of review applicable to the impugned portion of the trial judge’s credibility or reliability assessment. The standard of review will be correctness if the error alleged is a recognized error of law, such as reliance on myths and stereotypes about sexual assault complainants, or improper and incorrect assumptions about accused persons contrary to fundamental principles such as the right to silence and the presumption of innocence. Stereotypes based in other forms of inequality of treatment that are analogous to myths and stereotypes about sexual assault complainants may also be recognized as errors of law in future cases. Testimonial assessments may also become vulnerable to correctness review for reasonable apprehension of bias, making a finding of fact for which there is no evidence, and improperly taking judicial notice.

                    Absent an error of law, the standard of review will be palpable and overriding error. The reviewing court must first determine whether the erroneous reliance on the assumption is palpable, such as where the assumption in question is obviously untrue on its face, or where it is untrue or inapplicable in light of the other accepted evidence or findings of fact. Once a palpable error has been identified, the reviewing court must also find that the erroneous reliance on the assumption was overriding, where it has affected the result or goes to the very core of the outcome of the case. If it cannot be shown that the error was palpable and overriding, a trial judge’s assessment of credibility or reliability will be entitled to deference and there will be no basis for appellate intervention.

                    In K and T’s cases, the Court of Appeal erred in using the rule against ungrounded common‑sense assumptions and reviewing the alleged improper generalizations on a correctness standard. The trial decisions in both cases should have been reviewed for palpable and overriding error. In K’s case, the Court of Appeal erred in concluding that the trial judge relied on speculative reasoning in accepting the complainant’s evidence based on his observation that it is extremely unlikely that a woman would be mistaken about the feeling of penile‑vaginal penetration. Viewing the reasons as a whole and in context, the trial judge did not reject the defence theory because of an assumption that no woman would be mistaken, but rather because he accepted the complainant’s testimony that she, herself, was not mistaken. The trial judge’s conclusion was grounded in his assessment of the complainant’s testimony and no palpable and overriding errors were made. In T’s case, the Court of Appeal erred in concluding that the trial judge’s assessment of the accused and the complainant’s credibility was fatally affected by three material unfounded assumptions about normal behaviour. The first two assumptions were not assumptions but statements that reflected the trial judge’s reasoning process and findings of fact. The third assumption was truly an assumption and one that was palpably incorrect. However, it was not overriding, as it did not affect the core of the trial judge’s finding of guilt.

                    Per Rowe J.: Generalized expectations based on common sense and human experience play a necessary role in the judicial fact‑finding process; however, reliance on generalized expectations in a criminal proceeding is not without limits, as some expectations may not be accurate or reliable predictors of general human behaviour. The proliferation of appellate jurisprudence identifying concerns about the limits of this exercise points to the need for guidance in the form of a clear and consistent framework for appellate review. This framework is composed of three questions that an appellate court should ask when reviewing for potential legal error in a trial judge’s reliance on generalized expectations in the fact‑finding process: (1) Did the trial judge rely on a generalized expectation in their reasoning process? (2) If the trial judge relied on a generalized expectation, was the expectation reasonable? (3) Did the trial judge rely on a generalized expectation as itself a conclusive and indisputable fact?

                    The first question asks whether the trial judge relied on a generalized expectation in their reasoning process. If the appellate court determines that the judge relied on a generalized expectation, the analysis will proceed to the second question. Where a judge has not relied on any generalized expectation, for example where the judge assessed the evidence with reference to other accepted evidence or facts from the trial, the review for potential error under the framework ends. At this stage of the analysis, the appellate court is not identifying any error. It is not an error of law per se for a judge to rely on a generalized expectation as a logical benchmark to assess the evidence; rather, it is a well‑recognized and necessary part of the judicial fact‑finding process. In answering this first question, the appellate court is simply determining what the trial judge really decided, why the judge decided that way, and whether there is a basis for further scrutiny. This is a highly case‑dependent inquiry — an appellate court must assess whether the reasons, read as a whole and in context of the live issues at trial, explain what the trial judge decided and why they decided that way in a manner that permits effective appellate review. Appellate courts must not finely parse the trial judge’s reasons in search for error.

                    If the trial judge relied on a generalized expectation, the second question then asks whether the expectation was reasonable. It is an error of law for a judge to rely on an unreasonable generalized expectation to assess the evidence. Unreasonable generalized expectations masquerading as common sense or collective human experience are not a legitimate basis on which to assess and understand the evidence in a criminal trial. A standard of reasonableness imparts a measure of objectivity and community consensus in shaping the boundaries of a judge’s reliance on common sense and human experience to make decisions, ensuring that triers of fact do not rely on generalized expectations that are inaccurate or unreliable. It also recognizes that as society evolves, its understanding of what is reasonable may change. Public confidence in the administration of justice and the judicial fact‑finding process require that appellate courts be able to intervene where trial judges employ generalized expectations that are not a reasonably accurate reflection of what is true in most circumstances and are not a reliable benchmark to assess the evidence. However, where a trial judge relies on a reasonable generalized expectation, an appellate court may not interfere with the judge’s assessment of the evidence based on that expectation just because it would have come to a different conclusion.

                    Appellate review for legal error should extend to the reasonableness of any generalized expectation relied on by the trial judge. The categories of generalized expectations that may not, as a matter of law, be relied on should not be limited to what might be considered myths or stereotypes or to the context of sexual assault trials. Such a narrow approach creates artificial distinctions in the law, unduly limits the important role of appellate courts in the criminal justice system, and risks undermining public confidence in the administration of justice. On a conceptual level, the underlying rationale for reliance on a generalized expectation is the same, regardless of whether the generalized expectation can be considered a myth, a stereotype, or something else. There is no principled basis to separate particular categories of generalized expectations for distinct treatment. Such an exercise invites artificial distinctions in the law with no clear boundaries and fails to account for generalized expectations that cannot accurately be classified as myths or stereotypes about complainants in sexual assault cases yet would nonetheless be unacceptable to rely on in a criminal trial.

                    All reliance on unreasonable generalized expectations should be recognized as an error or law. Support for this is found in the relevant policy reasons for why appellate courts defer to trial judges’ factual findings on one hand, and the wide discretion for appellate intervention on questions of law on the other hand. Appellate courts defer to trial judges on factual matters for at least three overarching policy reasons: (1) to limit the number, length, and cost of appeals; (2) to promote the autonomy and integrity of trial proceedings; and (3) to recognize the expertise of the trial judge and the judge’s advantageous position to assess the evidence. Reviewing unreasonable expectations on a standard of correctness would have a limited impact on those policy considerations. Concerns about preserving public resources or promoting the autonomy and integrity of trials are secondary to the rights of the accused in a criminal proceeding, in light of the interests at stake and Parliament’s decision to provide broad access to a first level of appeal. It is the recognition of the trial judge’s expertise and advantageous position that forms the central basis for appellate deference on factual matters in the context of criminal proceedings; however, trial judges are not more experienced or in a more advantageous position than appellate judges in identifying a reasonable generalized expectation based on common sense or human experience, as the basis for the general reliability of such expectations is that they are common or the shared experience of the entire community.

                    The primary role of appellate courts is to delineate and refine legal rules and ensure their universal application. They therefore maintain a broad scope of review on questions of law, which require clear and consistent answers in order to maintain public confidence in the administration of justice. Any approach to appellate review that dilutes the important role of appellate courts would do significant damage to the criminal justice system. The characterization of an issue as a question of law is also significant to the grounds for appellate jurisdiction in criminal matters, especially the limited grounds for Crown appeals from acquittals. If the reasonableness of a generalized expectation were treated as a question of fact, then the Crown would not be able to appeal from acquittals that rest on generalized expectations that do not amount to myths or stereotypes, no matter how unreasonable or contrary to society’s collective expectations. The need to preserve public confidence in the administration of justice mandates that verdicts in criminal cases not be founded on assumptions that are not reasonably accurate reflections of what is true in most circumstances. The correctness standard requires setting out expressly an alternative line of reasoning and demonstrating why it should be followed. Properly applied, correctness demands greater conceptual clarity and analytical rigour. By contrast, the palpable and overriding standard of review can be misused to generate a decision‑making black box that facilitates ad hoc decision‑making, whereby if the Court of Appeal agrees with what the trial judge has done, it shows deference, but if it prefers another outcome, it labels the trial judge’s decision as a palpable and overriding error and substitutes its preferred outcome.

                    Finally, the third question asks whether the judge relied on a generalized expectation as itself a conclusive and indisputable fact. Trial judges have considerable latitude to rely on reasonable general expectations as a logical benchmark in assessing the evidence, and an appellate court may not interfere with the judge’s assessment of the evidence just because it would have come to a different conclusion. However, there is an important limit on the use of even a reasonable generalized expectation: the trial judge cannot treat the generalized expectation as itself a conclusive and indisputable fact, such that the judge ignores or forecloses their mind to the evidence. This is because people may always act contrary to a generalized expectation of what common sense or human experience would ordinarily anticipate, and therefore it is the trial judge’s duty to determine on the evidence what really happened. Generalized expectations based on human experience and common sense are only one consideration, which assist with interpreting the evidence, but the focus must remain on the evidence. It is an error of law for a judge to fail to consider all of the evidence on the ultimate issue of guilt or innocence, or to make a factual conclusion in the absence of evidence. Where a judge relies on a generalized expectation as itself a factual conclusion, what the judge is really doing is taking judicial notice, which is subject to a stringent test — a standard that will rarely, if ever, be met by a generalized expectation about people due to the variability of human experience and behaviour. And where a factual conclusion is based neither on the evidence, nor on judicial notice, then it is speculation, which is an error of law.

                    Applying these principles to the cases at hand, the appeals should be allowed and the convictions restored. In K’s case, under the framework’s first question, the trial judge relied on a generalized expectation about the likelihood of a woman being mistaken about the feeling of vaginal penetration. Under the framework’s second question, it is a reasonable generalized expectation about general human perception that a woman is unlikely to be mistaken about the feeling of vaginal penetration, as a sexual act of this nature would have a profound and traumatic impact on the bodily integrity of an individual, and ordinary people would not generally require special knowledge to assess this sort of evidence. Under the framework’s third question, the trial judge did not treat this as an indisputable fact but instead used it as a benchmark to assess the complainant’s evidence in light of the totality of the evidence. There was therefore no basis for the Court of Appeal to intervene.

                    In T’s case, the Court of Appeal wrongly identified the trial judge as having relied on a generalized expectation regarding the first two assumptions, where the latter was actually assessing the whole of the evidence; this led the Court of Appeal to move beyond its role and reweigh the evidence. In respect of the third assumption, the trial judge did err by relying on an unreasonable expectation about how people ordinarily behave after a consensual sexual encounter. Under the framework’s second question, it is unreasonable to expect any logical connection between an individual waiting for their sexual partner to enter a home and the consensual or non‑consensual nature of the preceding encounter. Nevertheless, this error of law was so harmless or minor that it could not have had any impact on the verdict, because the judge’s verdict clearly turned on her favourable assessment of the complainant’s credibility, as well as her complete rejection of T’s testimony as inconsistent with the events and contrived to explain away the complainant’s injuries.

Cases Cited

By Martin J.

                    Overruled: R. v. Roth, 2020 BCCA 240, 66 C.R. (7th) 107; R. v. Perkins, 2007 ONCA 585, 223 C.C.C. (3d) 289; R. v. Cepic, 2019 ONCA 541, 376 C.C.C. (3d) 286; R. v. J.C., 2021 ONCA 131, 401 C.C.C. (3d) 433; referred to: R. v. W. (D.), [1991] 1 S.C.R. 742; R. v. Vuradin, 2013 SCC 38, [2013] 2 S.C.R. 639; R. v. D. (J.J.R.) (2006), 215 C.C.C. (3d) 252; R. v. G.F., 2021 SCC 20, [2021] 1 S.C.R. 801; R. v. Mills, [1999] 3 S.C.R. 668; R. v. Darrach, 2000 SCC 46, [2000] 2 S.C.R. 443; R. v. J.J., 2022 SCC 28; R. v. A. (A.B.), 2019 ONCA 124, 145 O.R. (3d) 634; R. v. A.R.J.D., 2018 SCC 6, [2018] 1 S.C.R. 218; R. v. C.M.M., 2020 BCCA 56; R. v. Kodwat, 2017 YKCA 11; R. v. Thompson, 2019 BCCA 1, 370 C.C.C. (3d) 354; R. v. T.L., 2020 NUCA 10, 393 C.C.C. (3d) 195; R. v. Seaboyer, [1991] 2 S.C.R. 577; R. v. D.D., 2000 SCC 43, [2000] 2 S.C.R. 275; Kribs v. The Queen, [1960] S.C.R. 400; Timm v. The Queen, [1981] 2 S.C.R. 315; R. v. Friesen, 2020 SCC 9, [2020] 1 S.C.R. 424; R. v. Osolin, [1993] 4 S.C.R. 595; R. v. A.G., 2000 SCC 17, [2000] 1 S.C.R. 439; R. v. McCraw, [1991] 3 S.C.R. 72; R. v. Esau, [1997] 2 S.C.R. 777; R. v. Ewanchuk, [1999] 1 S.C.R. 330; R. v. Cinous, 2002 SCC 29, [2002] 2 S.C.R. 3; R. v. Kirkpatrick, 2022 SCC 33; R. v. Goldfinch, 2019 SCC 38, [2019] 3 S.C.R. 3; R. v. M. (M.L.), [1994] 2 S.C.R. 3; R. v. Find, 2001 SCC 32, [2001] 1 S.C.R. 863; R. v. Barton, 2019 SCC 33, [2019] 2 S.C.R. 579; R. v. Lacombe, 2019 ONCA 938, 383 C.C.C. (3d) 114; R. v. O’Connor, [1995] 4 S.C.R. 411; R. v. G. (C.M.), 2016 ABQB 368, 41 Alta. L.R. (6th) 374; R. v. F. (W.J.), [1999] 3 S.C.R. 569; R. v. Oakes, [1986] 1 S.C.R. 103; R. v. Noble, [1997] 1 S.C.R. 874; R. v. Laboucan, 2010 SCC 12, [2010] 1 S.C.R. 397; R. v. Van, 2009 SCC 22, [2009] 1 S.C.R. 716; R. v. J.H.S., 2008 SCC 30, [2008] 2 S.C.R. 152; R. v. H. (C.W.) (1991), 68 C.C.C. (3d) 146; R. v. S. (W.D.), [1994] 3 S.C.R. 521; R. v. Avetysan, 2000 SCC 56, [2000] 2 S.C.R. 745; R. v. Villaroman, 2016 SCC 33, [2016] 1 S.C.R. 1000; R. v. Esquivel‑Benitez, 2020 ONCA 160, 61 C.R. (7th) 326; Wild v. The Queen, [1971] S.C.R. 101; Rousseau v. The Queen, [1985] 2 S.C.R. 38; R. v. B. (G.), [1990] 2 S.C.R. 57; R. v. Clark, 2015 BCCA 488, 407 D.L.R. (4th) 610, aff’d 2017 SCC 3, [2017] 1 S.C.R. 86; R. v. Drydgen, 2021 BCCA 125; R. v. Petrolo, 2021 ONCA 498, 156 O.R. (3d) 321; J.P. v. R., 2022 QCCA 104; R. v. Calnen, 2019 SCC 6, [2019] 1 S.C.R. 301; R. v. Munoz (2006), 86 O.R. (3d) 134; R. v. Delmas, 2020 ABCA 152, 452 D.L.R. (4th) 375, aff’d 2020 SCC 39, [2020] 3 S.C.R. 780; R. v. R.R., 2018 ABCA 287, 366 C.C.C. (3d) 293; R. v. S. (R.D.), [1997] 3 S.C.R. 484; R. v. Kiss, 2018 ONCA 184; R. v. Adebogun, 2021 SKCA 136, [2022] 1 W.W.R. 187; R. v. Kontzamanis, 2011 BCCA 184; R. v. George, 2017 SCC 38, [2017] 1 S.C.R. 1021; R. v. Gagnon, 2006 SCC 17, [2006] 1 S.C.R. 621; Maxwell v. The Queen, [1979] 2 S.C.R. 1072; R. v. Norman (1993), 16 O.R. (3d) 295; Faryna v. Chorny, [1952] 2 D.L.R. 354; R. v. B. (K.G.), [1993] 1 S.C.R. 740; R. v. N.S., 2012 SCC 72, [2012] 3 S.C.R. 726; R. v. J.A.A., 2011 SCC 17, [2011] 1 S.C.R. 628; R. v. Rhayel, 2015 ONCA 377, 324 C.C.C. (3d) 362; R. v. Pelletier (1995), 165 A.R. 138; R. v. Virk, 2015 BCSC 981; R. v. R.E.M., 2008 SCC 51, [2008] 3 S.C.R. 3; Schwartz v. Canada, [1996] 1 S.C.R. 254; H.L. v. Canada (Attorney General), 2005 SCC 25, [2005] 1 S.C.R. 401; Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235; R. v. Chung, 2020 SCC 8, [2020] 1 S.C.R. 405; R. v. Morrissey (1995), 22 O.R. (3d) 514; R. v. Pastro, 2021 BCCA 149, 71 C.R. (7th) 296; R. v. Greif, 2021 BCCA 187; R. v. Al‑Rawi, 2021 NSCA 86, 410 C.C.C. (3d) 385; R. v. K.B.W., 2022 SKCA 8; R. v. L.L., 2022 ONCA 50; R. v. Lapierre, 2022 NSCA 12; Kritik‑Langer v. R., 2022 QCCA 657; R. v. Kavanagh, 2022 BCCA 225; R. v. D.B., 2022 SKCA 76, 415 C.C.C. (3d) 455; R. v. S.A., 2022 ONCA 642; R. v. S.M., 2023 ONCA 417; South Yukon Forest Corp. v. Canada, 2012 FCA 165, 431 N.R. 286; Benhaim v. St‑Germain, 2016 SCC 48, [2016] 2 S.C.R. 352; Salomon v. Matte‑Thompson, 2019 SCC 14, [2019] 1 S.C.R. 729; R. v. J.M.H., 2011 SCC 45, [2011] 3 S.C.R. 197; Schuldt v. The Queen, [1985] 2 S.C.R. 592; R. v. Poperechny, 2020 MBCA 81, 396 C.C.C. (3d) 478; R. v. Clark, 2005 SCC 2, [2005] 1 S.C.R. 6; R. v. Corbett, [1988] 1 S.C.R. 670; R. v. White, 2011 SCC 13, [2011] 1 S.C.R. 433.

By Rowe J.

                    Overruled: R. v. J.C., 2021 ONCA 131, 401 C.C.C. (3d) 433; considered: R. v. Quartey, 2018 ABCA 12, 430 D.L.R. (4th) 381, aff’d 2018 SCC 59, [2018] 3 S.C.R. 687; R. v. L. (J.), 2018 ONCA 756, 143 O.R. (3d) 170; R. v. Roth, 2020 BCCA 240, 66 C.R. (7th) 107; referred to: R. v. Kodwat, 2017 YKCA 11; R. v. A.R.D., 2017 ABCA 237, 422 D.L.R. (4th) 471, aff’d 2018 SCC 6, [2018] 1 S.C.R. 218; R. v. Kiss, 2018 ONCA 184; R. v. Paulos, 2018 ABCA 433, 79 Alta. L.R. (6th) 33; R. v. A. (A.B.), 2019 ONCA 124, 145 O.R. (3d) 634; R. v. F.B.P., 2019 ONCA 157; R. v. Cepic, 2019 ONCA 541, 376 C.C.C. (3d) 286; R. v. Pilkington, 2019 BCCA 374; R. v. Percy, 2020 NSCA 11, 61 C.R. (7th) 7; R. v. Delmas, 2020 ABCA 152, 452 D.L.R. (4th) 375, aff’d 2020 SCC 39, [2020] 3 S.C.R. 780; R. v. Pastro, 2021 BCCA 149, 71 C.R. (7th) 296; R. v. Perkins, 2007 ONCA 585, 223 C.C.C. (3d) 289; R. v. Find, 2001 SCC 32, [2001] 1 S.C.R. 863; R. v. François, [1994] 2 S.C.R. 827; R. v. D.A.I., 2012 SCC 5, [2012] 1 S.C.R. 149; R. v. H.C., 2009 ONCA 56, 244 O.A.C. 288; R. v. Gagnon, 2006 SCC 17, [2006] 1 S.C.R. 621; R. v. Cinous, 2002 SCC 29, [2002] 2 S.C.R. 3; R. v. Villaroman, 2016 SCC 33, [2016] 1 S.C.R. 1000; R. v. Gibson, 2021 ONCA 530, 157 O.R. (3d) 597; R. v. Chanmany, 2016 ONCA 576, 338 C.C.C. (3d) 578; Lampard v. The Queen, [1969] S.C.R. 373; White Burgess Langille Inman v. Abbott and Haliburton Co., 2015 SCC 23, [2015] 2 S.C.R. 182; Graat v. The Queen, [1982] 2 S.C.R. 819; R. v. Mohan, [1994] 2 S.C.R. 9; R. v. Morrissey (1995), 22 O.R. (3d) 514; Canadian Pacific Railway Co. v. Murray, [1932] S.C.R. 112; R. v. Béland, [1987] 2 S.C.R. 398; R. v. S. (R.D.), [1997] 3 S.C.R. 484; R. v. Marquard, [1993] 4 S.C.R. 223; R. v. Walle, 2012 SCC 41, [2012] 2 S.C.R. 438; R. v. Jacquard, [1997] 1 S.C.R. 314; R. v. O’Brien, [1978] 1 S.C.R. 591; R. v. Chouhan, 2021 SCC 26, [2021] 2 S.C.R. 136; R. v. G.F., 2021 SCC 20, [2021] 1 S.C.R. 801; R. v. Chung, 2020 SCC 8, [2020] 1 S.C.R. 405; R. v. Sarrazin, 2011 SCC 54, [2011] 3 S.C.R. 505; R. v. Van, 2009 SCC 22, [2009] 1 S.C.R. 716; R. v. Khan, 2001 SCC 86, [2001] 3 S.C.R. 823; R. v. Samaniego, 2022 SCC 9; R. v. Graveline, 2006 SCC 16, [2006] 1 S.C.R. 609; R. v. Morin, [1988] 2 S.C.R. 345; R. v. Biniaris, 2000 SCC 15, [2000] 1 S.C.R. 381; R. v. Lohrer, 2004 SCC 80, [2004] 3 S.C.R. 732; R. v. Sinclair, 2011 SCC 40, [2011] 3 S.C.R. 3; R. v. Shepherd, 2009 SCC 35, [2009] 2 S.C.R. 527; R. v. R.R., 2018 ABCA 287, 366 C.C.C. (3d) 293; R. v. White, 2011 SCC 13, [2011] 1 S.C.R. 433; R. v. A.G., 2000 SCC 17, [2000] 1 S.C.R. 439; R. v. D.D., 2000 SCC 43, [2000] 2 S.C.R. 275; R. v. Mills, [1999] 3 S.C.R. 668; R. v. Osolin, [1993] 4 S.C.R. 595; R. v. J.M.H., 2011 SCC 45, [2011] 3 S.C.R. 197; Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235; R. v. Spencer, 2007 SCC 11, [2007] 1 S.C.R. 500; R. v. Beaudry, 2007 SCC 5, [2007] 1 S.C.R. 190; Fanjoy v. The Queen, [1985] 2 S.C.R. 233; R. v. Ewanchuk, [1999] 1 S.C.R. 330; R. v. J.C., 2018 ONSC 5547; Schuldt v. The Queen, [1985] 2 S.C.R. 592; R. v. Morin, [1992] 3 S.C.R. 286; R. v. Walker, 2008 SCC 34, [2008] 2 S.C.R. 245.

Statutes and Regulations Cited

Act to amend the Criminal Code in relation to sexual offences and other offences against the person and to amend certain other Acts in relation thereto or in consequence thereof, S.C. 1980‑81‑82‑83, c. 125, ss. 5, 19.

Act to amend the Criminal Code (sexual assault), S.C. 1992, c. 38, preamble.

Canadian Charter of Rights and Freedoms , s. 11(c) , (d).

Criminal Code , R.S.C. 1985, c. C‑46, ss. 273.1(1)  “consent”, (2), 273.2, 274, 275, 276, 278.1 to 278.91, 686(1)(a), (b)(iii).

Criminal Code, S.C. 1953‑54, c. 51, s. 131(1).

Authors Cited

Benedet, Janine. “Judicial Misconduct in the Sexual Assault Trial” (2019), 52 U.B.C. L. Rev. 1.

Benedet, Janine, and Isabel Grant. “Hearing the Sexual Assault Complaints of Women with Mental Disabilities: Evidentiary and Procedural Issues” (2007), 52 McGill L.J. 515.

Cochran, Patricia. Common Sense and Legal Judgment: Community Knowledge, Political Power, and Rhetorical Practice. Montréal: McGill‑Queen’s University Press, 2017.

Coughlan, Steve. Criminal Procedure, 4th ed. Toronto: Irwin Law, 2020.

Dufraimont, Lisa. “Current Complications in the Law on Myths and Stereotypes” (2021), 99 Can. Bar Rev. 536.

Dufraimont, Lisa. “Myth, Inference and Evidence in Sexual Assault Trials” (2019), 44 Queen’s L.J. 316.

Ehrlich, Susan. “Perpetuating — and Resisting — Rape Myths in Trial Discourse”, in Elizabeth A. Sheehy, ed., Sexual Assault in Canada: Law, Legal Practice and Women’s Activism. Ottawa: University of Ottawa Press, 2012, 389.

Hill, S. Casey, David M. Tanovich and Louis P. Strezos. McWilliams’ Canadian Criminal Evidence, 5th ed. Toronto: Thomson Reuters, 2013 (loose‑leaf updated December 2023, release 5).

Oxford English Dictionary (online: www.oed.com), “myth”.

Paciocco, David M., Palma Paciocco and Lee Stuesser. The Law of Evidence, 8th ed. Toronto: Irwin Law, 2020.

Koshan, Jennifer. “Marriage and Advance Consent to Sex: A Feminist Judgment in R v JA” (2016), 6 Oñati Socio‑legal Series 1377.

Lederman, Sidney N., Michelle K. Fuerst and Hamish C. Stewart. Sopinka, Lederman & Bryant: The Law of Evidence in Canada, 6th ed. Toronto: LexisNexis, 2022.

Randall, Melanie. “Sexual Assault Law, Credibility, and ‘Ideal Victims’: Consent, Resistance, and Victim Blaming” (2010), 22 C.J.W.L. 397.

Tanovich, David M. “Regulating Inductive Reasoning in Sexual Assault Cases”, in Benjamin L. Berger, Emma Cunliffe and James Stribopoulos, eds., To Ensure that Justice is Done: Essays in Memory of Marc Rosenberg. Toronto: Thomson Reuters, 2017, 73.

Watt, David. Watt’s Manual of Criminal Evidence. Toronto: Thomson Reuters, 2023.

                    APPEAL from a judgment of the British Columbia Court of Appeal (Tysoe, Stromberg‑Stein and Marchand JJ.A.), 2022 BCCA 18, [2022] B.C.J. No. 90 (Lexis), 2022 CarswellBC 102 (WL), setting aside the conviction entered by Tammen J., 2020 BCSC 1480, [2020] B.C.J. No. 2418 (Lexis), 2020 CarswellBC 3779 (WL), and ordering a new trial. Appeal allowed.

                    APPEAL from a judgment of the British Columbia Court of Appeal (Newbury, Willcock and Voith JJ.A.), 2022 BCCA 345, 419 C.C.C. (3d) 187, 84 C.R. (7th) 314, [2022] B.C.J. No. 1943 (Lexis), 2022 CarswellBC 2858 (WL), setting aside the conviction entered by Phillips Prov. Ct. J., 2020 BCPC 306, [2020] B.C.J. No. 2446 (Lexis), 2020 CarswellBC 3826 (WL), and ordering a new trial. Appeal allowed.

                    Susanne Elliott, Christie Lusk and Lauren Chu, for the appellant.

                    Brent R. Anderson and Christopher Johnson, K.C., for the respondent Christopher James Kruk.

                    Richard S. Fowler, K.C., and Eric Purtzki, for the respondent Edwin Tsang.

                    Sarah Clive, for the intervener the Attorney General of Alberta.

                    Gregory P. DelBigio, K.C., and Daniel J. Song, K.C., for the intervener the Independent Criminal Defence Advocacy Society.

                    Breana Vandebeek, for the intervener the Criminal Lawyers’ Association (Ontario).

                    Mark Iyengar and Benjamin Reedijk, for the intervener the Trial Lawyers Association of British Columbia.

                    Hugo Caissy and Myralie Roussin, for the intervener Association québécoise des avocats et avocates de la défense.

                    Megan Stephens, Humera Jabir and Roxana Parsa, for the interveners the West Coast Legal Education and Action Fund Association and the Women’s Legal Education and Action Fund Inc.

                   The judgment of Wagner C.J. and Côté, Martin, Kasirer, Jamal and O’Bonsawin JJ. was delivered by

                   Martin J. —

I.               Introduction

[1]                             These appeals in two sexual assault matters concern the standard for appellate intervention with respect to a trial judge’s credibility and reliability findings in a criminal trial and the appropriate role of common sense when assessing the evidence of witnesses. The respondents ask this Court to recognize a novel rule referred to as the “rule against ungrounded common-sense assumptions”. A breach of this proposed rule would provide a new, stand-alone basis for correctness review of credibility and reliability assessments whenever an appellate court determines that a trial judge has relied on a common-sense assumption that was not grounded in the evidence. This significant departure from established standards of review in respect of credibility and reliability assessments in criminal cases has been applied by some appellate courts — often in sexual assault cases that turn on the competing accounts of the accused and the complainant.

[2]                             For the reasons provided below, no such change to the law is warranted, and I decline to recognize the rule against ungrounded common-sense assumptions as giving rise to an error of law. The current standards under which appellate courts review trial judgments are well-designed, long-established, and promote the fair assessment of testimony. There is no need to fashion a new rule of law against any assumption not supported by particular evidence in the record to strive for what existing rules already accomplish. Furthermore, the proposed rule is not a coherent extension of existing errors of law pertaining to myths and stereotypes against sexual assault complainants. Adopting it would undercut the functional and flexible approach to appellate intervention and create mischief across the entire criminal law.

[3]                             The faulty use of common-sense assumptions in criminal trials will continue to be controlled by existing standards of review and rules of evidence. In some cases, a trial judge’s use of common sense will be vulnerable to appellate review because it discloses recognized errors of law. Otherwise, like with other factual findings, credibility and reliability assessments — and any reliance on the common-sense assumptions inherent within them — will be reviewable only for palpable and overriding error. This standard is better equipped to the task than the new error of law the respondents propose.

[4]                             In both cases before us, the Court of Appeal for British Columbia overturned the sexual assault convictions on the basis of alleged errors of law in the trial judges’ credibility and reliability assessments. Using the rule against ungrounded common-sense assumptions, the Court of Appeal found that the trial judges erred in law by making assumptions about human behaviour not grounded in the evidence. Having rejected this new error of law, I would assess the trial judges’ findings using the proper standard of palpable and overriding error. I conclude that they made no such errors in their credibility and reliability findings. In the result, I would allow both appeals and restore the convictions.

II.            Background

A.           Mr. Kruk

[5]                             The respondent Christopher James Kruk was charged with sexual assault based on an allegation of non-consensual penile-vaginal penetration. The main issue at trial was whether the sexual activity occurred, which depended almost entirely upon the trial judge’s assessment of the credibility of Mr. Kruk and the reliability of the complainant.

[6]                             Mr. Kruk found the complainant intoxicated, lost, and distressed one night in downtown Vancouver. He decided to take her to his house, and connected with the complainant’s parents by phone. The complainant did not remember much of what happened next, but she testified that she woke up with her pants off and Mr. Kruk on top of her with his penis inside her vagina. Mr. Kruk denied this, testifying that after the complainant spilled a glass of water on herself and went to his room to change, he found her passed out on his bed with her pants around her ankles. He said that when he tried to wake her, she was startled and must have assumed the worst, because she kicked off her pants and began running around the house in a panic.

(1)          Trial Decision, 2020 BCSC 1480

[7]                             The trial judge found Mr. Kruk guilty of sexual assault. He did not believe Mr. Kruk and rejected his evidence on multiple points because of material inconsistencies in his testimony. Specifically, the trial judge rejected Mr. Kruk’s explanation of the complainant’s state of undress when she awoke, and “flatly rejected” his claim that he contacted the complainant’s family to get her home safely. In turn, the trial judge considered the complainant’s account to be largely unreliable due to her intoxication and “massive gaps in her memory” (para. 48 (CanLII)). However, the trial judge accepted her evidence on the core of the sexual assault allegation — that she had felt Mr. Kruk’s penis inside of her vagina:

[The complainant’s] evidence is devoid of detail, yet she claims to be certain that she was not mistaken. She said she felt [Mr. Kruk’s] penis inside her and she knew what she was feeling. In short, her tactile sense was engaged. It is extremely unlikely that a woman would be mistaken about that feeling. [Emphasis added; para. 68.]

[8]                             The trial judge also held there was important circumstantial evidence consistent with a sexual encounter, providing further reason to accept the complainant’s account and reject Mr. Kruk’s. At one point while the complainant was in Mr. Kruk’s bedroom, he had taken off all his clothes and put on swim trunks. The complainant’s pants had been completely removed, and the trial judge considered the notion that the complainant had done so herself to be “fanciful” given her state of intoxication. Finally, phone records admitted at trial revealed that after their initial conversation with Mr. Kruk, the complainant’s parents had repeatedly called and texted and he failed to respond. In the overall circumstances, the trial judge considered that Mr. Kruk’s failure to contact the complainant’s parents was consistent with an intention to take advantage of her sexually while she was in an extremely vulnerable state.

(2)          British Columbia Court of Appeal Decision, 2022 BCCA 18

[9]                             The Court of Appeal unanimously overturned Mr. Kruk’s conviction and ordered a new trial. The court observed that although trial judges can rely on personal life experiences to assess the credibility of witnesses, they “must be careful . . . to ‘avoid speculative reasoning that invokes « common sense » assumptions not grounded in the evidence’” (para. 41 (CanLII), quoting R. v. Roth, 2020 BCCA 240, 66 C.R. (7th) 107, at para. 65). The trial judge’s conclusion regarding the unlikelihood of any complainant (“a woman”) being mistaken about the feeling of penile‑vaginal penetration relied on speculative reasoning and was not the proper subject of judicial notice, engaging “questions of neurology (the operation of the body’s sensory system), physiology (the impact of alcohol on perception, memory and the body’s sensory system) and psychiatry (the impact of alcohol and/or trauma on perception and memory)” (paras. 65 and 67). Since the trial judge’s finding regarding penetration was considered to be the primary reason he had accepted the complainant’s evidence, the court held that the trial judge’s reasoning had been fatally affected by a legal error.

B.            Mr. Tsang

[10]                         The respondent Edwin Tsang was charged with sexual assault based on alleged non-consensual fellatio, digital vaginal and anal penetration, and penile-vaginal penetration. Although the parties’ accounts of what sexual activity had happened also diverged, the main issue was consent, which primarily hinged on the credibility of both the complainant and the accused.

[11]                         The complainant and Mr. Tsang met at a music festival she attended with her friend, and the three subsequently went to an after-party at a club in downtown Vancouver, where they danced and drank. Mr. Tsang then offered to drive the complainant home, and on the way, they stopped in a parking lot. They got into the backseat and the complainant consented to kissing Mr. Tsang. The complainant’s and Mr. Tsang’s testimonies diverged significantly at this point. The complainant testified that Mr. Tsang, all without her consent, forced her to perform fellatio on him, digitally penetrated her vagina and anus, and forced her to have vaginal intercourse. Mr. Tsang testified that they engaged in consensual, rough oral sex and digital penetration, and that the complainant had initiated and was an active participant during the sexual encounter. He also denied having vaginal intercourse or engaging in anal penetration.

[12]                         The morning after the alleged assault, the complainant was examined by a sexual assault physician examiner, who also testified as an expert at trial. The examiner documented extensive injuries to the complainant’s genital area, including a tear in her genital opening only typically seen after childbirth.

(1)          Trial Decision, 2020 BCPC 306

[13]                         The trial judge found Mr. Tsang guilty of sexual assault. She acknowledged that the complainant had consented to getting into the backseat of the car and to kissing, but rejected Mr. Tsang’s evidence that the subsequent sexual acts had been consensual, which she found to be at odds with the evidence she had accepted. The trial judge found that Mr. Tsang’s testimony sounded contrived, as if he were describing how he thought such an encounter might unfold rather than what had really happened. She also took issue with Mr. Tsang’s evidence that the complainant had initiated each step of the sexual encounter, finding instead that the accused had dominated and controlled the complainant, and “orchestrated” the events. By contrast, the judge found the complainant was credible and reliable and held that the Crown’s case, which relied heavily on her testimony, did not raise a reasonable doubt. The judge concluded that Mr. Tsang knew the complainant was not consenting as soon as he began forcing her to give him fellatio, and accordingly found him guilty of sexual assault.

[14]                         The trial judge made three factual findings that became crucial on appeal. First, it was not believable that the complainant had asked Mr. Tsang to spank her and was “gearing up for rough sex” (para. 126 (CanLII)). Second, Mr. Tsang’s explanation that they did not have vaginal intercourse because he could not find a condom was contrived and inconsistent with the level of control he had conveyed over all of their interactions that evening and in court. Third, Mr. Tsang’s rapid departure once he dropped the complainant off at her friend’s house was consistent with a sexual assault having occurred.

(2)          British Columbia Court of Appeal Decision, 2022 BCCA 345, 419 C.C.C. (3d) 187

[15]                         The Court of Appeal unanimously overturned Mr. Tsang’s conviction and ordered a new trial. The court found that the trial judge had made three assumptions about human behaviour that had impacted her assessment of the evidence: (1) a person would not ask to be spanked while engaging in sexual foreplay “out of the blue”; (2) a controlling person would not refrain from engaging in vaginal intercourse because they could not find a condom; and (3) a person would not abruptly and unceremoniously drive away from the person with whom they had engaged in consensual sex. Relying on the rule against ungrounded common-sense assumptions, these generalizations were not based in the evidence and engaged in speculative reasoning, and thus constituted reversible legal errors. The court also found these errors were material, as the judge would not necessarily have rejected Mr. Tsang’s evidence without erring in law.

III.         Analysis

[16]                         In both Mr. Kruk and Mr. Tsang’s cases, the British Columbia Court of Appeal reviewed the trial judges’ credibility and reliability findings on a standard of correctness. In line with this approach, the respondents now ask this Court to recognize the proposed rule against ungrounded common-sense assumptions as grounding an error of law.

[17]                         These appeals are part of a broader pattern in which special rules have been proposed for the assessment of credibility and reliability in sexual assault cases.^[1] The rule against ungrounded common-sense assumptions is before this Court for the first time, and the question of whether it should be adopted should be approached based on first principles relating to credibility and reliability assessments and existing standards of review. When the rule comes into play in sexual assault cases in particular, constitutional imperatives call for the consideration of the Charter  rights of both accused persons and complainants as well as the interests of society at large (R. v. Mills, [1999] 3 S.C.R. 668; R. v. Darrach, 2000 SCC 46, [2000] 2 S.C.R. 443; R. v. J.J., 2022 SCC 28). This approach is part of the obligation to strive towards “the fair, ethical and non-discriminatory adjudication of sexual assault cases” (D. M. Tanovich, “Regulating Inductive Reasoning in Sexual Assault Cases”, in B. L. Berger, E. Cunliffe and J. Stribopoulos, eds., To Ensure that Justice is Done: Essays in Memory of Marc Rosenberg (2017), 73, at p. 95).

[18]                         These reasons proceed as follows. First, I explore the origins and treatment of the proposed rule. Next, I explain why it should not be recognized as giving rise to an error of law. Finally, applying the correct legal principles to both cases on appeal, I find that no palpable and overriding errors were made by the trial judge in Mr. Kruk’s case, and that while the trial judge in Mr. Tsang’s case relied on a palpably wrong generalization, the error was not overriding.

A.           The Proposed Rule Against Ungrounded Common-Sense Assumptions

[19]                         The proposed rule against ungrounded common-sense assumptions originated in a trio of key cases — R. v. Perkins, 2007 ONCA 585, 223 C.C.C. (3d) 289; R. v. Cepic, 2019 ONCA 541, 376 C.C.C. (3d) 286; and Roth — and has been most clearly and recently articulated in R. v. J.C., 2021 ONCA 131, 401 C.C.C. (3d) 433. The Court of Appeal in Mr. Kruk’s case cited to the trio of cases and specifically quoted the rule as stated by the British Columbia Court of Appeal in Roth, while the Court of Appeal in Mr. Tsang’s case relied on the rule as stated in J.C.

[20]                         In Perkins, the Ontario Court of Appeal held that the trial judge’s finding that a “virile young man with a full erection bound on having a climax would not lose his erection” was not the use of acceptable common sense, but was instead a conclusion reached outside the evidence and beyond the proper scope of judicial notice (paras. 35-36; see also para. 40). In Cepic, the Ontario Court of Appeal held that the trial judge’s characterization of the accused’s testimony as “implausible” and “nonsensical” was untethered to the evidence and reflected erroneous assumptions about what a young woman would do and stereotypes about male aggression (paras. 19-27). Both Perkins and Cepic were subsequently invoked in Roth, at para. 65, in which the British Columbia Court of Appeal articulated the following rule: “. . . although judges are entitled to rely on their human experience in assessing the plausibility of a witness’s testimony, they must avoid speculative reasoning that invokes ‘common sense’ assumptions not grounded in the evidence . . . .” The court went on to find that the trial judge had erred in relying on the speculative assumption that the accused, “a fit and healthy young man who regularly worked out and trained as a power lifter” (para. 56), would not fall asleep during a cab ride, erroneously leading the trial judge to reject the accused’s narrative.

[21]                         In J.C., the trial judge had found the accused guilty of sexual assault and extortion, accepting the complainant’s testimony that the accused had threatened to post a sexual video recording of her online if she did not continue her sexual relationship with him. On appeal, the accused alleged that the trial judge made two errors in his credibility assessment. First, the accused alleged that the trial judge impermissibly used stereotype to reject his testimony about expressly seeking the complainant’s consent before engaging in specific sexual acts — the trial judge had held that the accused’s testimony on this point was “too perfect, too mechanical, too rehearsed, and too politically correct to be believed” (para. 4). Second, the accused alleged that the trial judge had erred in characterizing the defence’s theory that the complainant had fabricated the allegations to conceal cheating from her boyfriend as relying on stereotype, when in fact it had been grounded in the evidence.

[22]                         The Ontario Court of Appeal unanimously allowed the appeal, set aside the convictions, and ordered a new trial. The court theorized a new approach to credibility and reliability assessments in sexual assault cases that consisted of two separate but interrelated rules: the new “rule against ungrounded common-sense assumptions”, and the modification of an existing “rule against stereotypical inferences”, the latter of which prevents judges from relying on stereotypes about how sexual assault complainants or accused persons are expected to act in assessing their credibility (J.C., at para. 63, citing Roth, at para. 129; R. v. A. (A.B.), 2019 ONCA 124, 145 O.R. (3d) 634, at para. 5; Cepic, at paras. 14 and 24). The court considered that failing to abide by either rule was an error of law reviewable on a standard of correctness where the error played a material role in the impugned conclusion.

[23]                         The appeals before this Court are therefore part of a body of recent jurisprudence that seeks to transform how credibility and reliability findings in sexual assault cases are reviewed on appeal. This jurisprudence undertakes three significant legal innovations. First, it introduces a novel general rule against ungrounded common-sense assumptions. Second, it classifies any breach of this rule as an error of law. Third, it takes the existing error of law barring the use of myths and stereotypes about sexual assault complainants and transposes it into a separate, parallel error of law relating to all factual assumptions about all witnesses, including the accused.

[24]                         Mindful of the context in which many of these cases have been argued, at the core of this novel approach is an explicit analogy between the historic treatment of myths and stereotypes undermining the credibility of sexual assault complainants and the principles to be applied when assessing the testimony of accused persons in sexual assault cases. In J.C., the court considered that because reliance on myths and stereotypes against sexual assault complainants is an error of law, it would be “equally wrong to draw inferences from stereotypes about the way accused persons are expected to act” (paras. 63; see paras. 72-74, citing R. v. A.R.J.D., 2018 SCC 6, [2018] 1 S.C.R. 218; A. (A.B.)). Embracing the proposed rule, other courts have agreed that this established error of law pertaining to myths and stereotypes about sexual assault complainants must also extend to other types of stereotypes or factual assumptions that can work to undermine an accused’s testimony. Collectively, these cases advance the notion that, although myths and stereotypes have traditionally operated to prejudice complainants, these same myths and stereotypes can simultaneously prejudice the accused (R. v. C.M.M., 2020 BCCA 56, at para. 139 (CanLII); R. v. Kodwat, 2017 YKCA 11, at para. 41 (CanLII); R. v. Thompson, 2019 BCCA 1, 370 C.C.C. (3d) 354, at paras. 56-57; R. v. T.L., 2020 NUCA 10, 393 C.C.C. (3d) 195, at para. 35).

[25]                         The parallel treatment proffered in this line of cases is at the forefront of the cases on appeal. At para. 25, the Court of Appeal in Mr. Tsang’s case relied upon a lengthy extract from J.C. that included the idea that it is “equally wrong” to invoke this historical set of myths and stereotypes about complainants in assessing the testimony of accused persons. Going further, in Mr. Kruk’s case, the Court of Appeal expressly drew a parallel between the historic use of myths and stereotypes as applied against sexual assault complainants, and the use of improper assumptions more generally:

Historically, the issue of improper assumptions typically arose in the context of trial judges drawing adverse inferences about the credibility of sexual assault complainants based on discredited myths and stereotypes. The authorities are clear, however, that trial judges must not rely on assumptions and stereotypes in their assessment of testimony given by both sexual assault complainants and accused persons. [Emphasis added; para. 42.]

[26]                         The proposed rule against ungrounded common-sense assumptions thus represents a radical departure from how appellate courts have typically approached credibility and reliability assessments, especially in the context of sexual assault. Crucially, the rule transformed all factual findings that could be said to constitute “assumptions” — which, in the ordinary course, were previously subject to review for palpable and overriding error — into errors of law reviewable on a standard of correctness. The rule also enlarged the well-established prohibition against myths and stereotypes about complainants — a relatively narrow error of law targeted at the fair assessment of a discrete category of witness testimony in a particular subset of criminal trials — well beyond its original scope, presenting it as the doctrinal basis for the creation of a far broader error that applies to credibility assessments of any witness in all criminal trials.

B.            The Proposed Rule Against Ungrounded Common-Sense Assumptions Should Not Be Adopted

[27]                         For the following reasons, the proposed rule against ungrounded common-sense assumptions should not be recognized as giving rise to an error of law.

[28]                         First, in the sexual assault context, the rule disregards the distinct nature of myths and stereotypes about complainants, transforming all factual generalizations regardless of their nature into errors of law and imposing a false symmetry to the circumstances of accused persons. Second, the rule runs contrary to long-settled law on credibility and reliability assessments and existing standards of review, leading to unprecedented and undesirable consequences. 

(1)          The Proposed Rule Is Not a Logical Extension of the Prohibition Against Myths and Stereotypes About Sexual Assault Complainants

[29]                         As noted, the proposed rule against ungrounded common-sense assumptions builds on the well-established prohibition against the use of myths and stereotypes when assessing the testimony of sexual assault complainants. In the ordinary course, credibility and reliability assessments are reviewable for palpable and overriding error and are otherwise entitled to deference. However, if a trial judge relies on myths or stereotypes about sexual assault complainants in their reasons, this amounts to an error of law. The proponents of the rule against ungrounded common-sense assumptions argue that if reliance on myths and stereotypes about complainants is an error of law, then all unfounded and speculative assumptions about all witnesses, including the accused, should be treated in the same manner.

[30]                         With respect, this impulse towards symmetry and formally identical treatment is unwarranted. It reflects a misunderstanding of the distinct body of law associated with myths and stereotypes in sexual assault cases, which developed in a particular historical context to protect complainants alone. 

(a)           The History of Myths and Stereotypes Against Complainants

[31]                         The prohibition against myths and stereotypes that undermine the credibility of sexual assault complainants has a unique history and a specific remedial purpose: to remove discriminatory legal rules that contributed to the view that women, as a group, were less worthy of belief and did not deserve legal protection against sexual violence.

[32]                         In the past, multiple legal barriers operated to ensure that the testimony of sexual assault complainants — who, at the time, were almost exclusively women — was treated as inherently unreliable. The term “myths and stereotypes” was coined to describe how the exceptional procedural protections historically afforded to those accused of sexual assault discriminated against complainants and made sexual assault not only the most underreported crime, but one that was exceptionally difficult to prove in court. These myths and stereotypes, formerly embedded into the law, arose in relation to both credibility and consent, and made sexual assault inherently dissimilar to other crimes.

[33]                         Before 1983, “rape” was a main type of sexual offence: it criminalized non-consensual penetration of a penis into a vagina and was thus understood as a gendered crime committed by men against women. Under the express terms of the prohibition, a married woman could not be raped by her husband as she was deemed, by her status as a wife, to have forfeited her legal capacity to refuse unwanted sexual activity. The law thereby enforced the notion that certain relationships provided men with a right to women’s sexual availability (see, e.g., J. Koshan, “Marriage and Advance Consent to Sex: A Feminist Judgment in R v JA” (2016), 6 Oñati Socio-legal Series 1377, at p. 1387).

[34]                         Special evidentiary rules also governed the testimony of sexual assault complainants. The statutory corroboration requirement obliged judges to instruct juries that it was dangerous to convict the accused of a sexual offence based only on the testimony of a complainant (Criminal Code, S.C. 1953-54, c. 51, s. 131(1); corroboration requirements were abolished by An Act to amend the Criminal Code in relation to sexual offences and other offences against the person and to amend certain other Acts in relation thereto or in consequence thereof, S.C. 1980-81-82-83, c. 125, ss. 5 and 19; see also Criminal Code , R.S.C. 1985, c. C-46  (“Code ”), s. 274; R. v. Seaboyer, [1991] 2 S.C.R. 577, at p. 676). The doctrine of recent complaint required immediate disclosure of a sexual assault to avoid an adverse credibility inference, meaning a complainant’s initial silence could be taken as “a virtual self-contradiction of her story” (R. v. D.D., 2000 SCC 43, [2000] 2 S.C.R. 275, at paras. 60-61, quoting Kribs v. The Queen, [1960] S.C.R. 400, at p. 405; see also Timm v. The Queen, [1981] 2 S.C.R. 315).

[35]                         By virtue of these rules, not only were women as a group seen as lacking credibility, there was a construct about the specific subset of women who could be believed. Negative social attitudes about women were often used to differentiate “real” rape victims from women suspected to be concocting false allegations out of self-interest or even revenge. Prejudicial beliefs about women who were Indigenous, racialized, persons with disabilities, or part of the 2SLGBTQI+ community also operated to socially prescribe ideals and norms about sexual assault victims. To the extent a particular complainant deviated from the characteristics of the idealized real rape victim, her credibility and entitlement to the law’s protection was diminished, sometimes to zero (see, e.g., J. Benedet, “Judicial Misconduct in the Sexual Assault Trial” (2019), 52 U.B.C. L. Rev. 1; S. Ehrlich, “Perpetuating — and Resisting — Rape Myths in Trial Discourse”, in E. A. Sheehy, ed., Sexual Assault in Canada: Law, Legal Practice and Women’s Activism (2012), 389; M. Randall, “Sexual Assault Law, Credibility, and ‘Ideal Victims’: Consent, Resistance, and Victim Blaming” (2010), 22 C.J.W.L. 397; J. Benedet and I. Grant, “Hearing the Sexual Assault Complaints of Women with Mental Disabilities: Evidentiary and Procedural Issues” (2007), 52 McGill L.J. 515).

[36]                         This Court has repeatedly recognized the prevalence of myths and stereotypes about sexual assault complainants, some of which include the following:

•         Genuine sexual assaults are perpetrated by strangers to the victim (Seaboyer, at p. 659, per L’Heureux-Dubé J., dissenting in part; R. v. Friesen, 2020 SCC 9, [2020] 1 S.C.R. 424, at para. 130, per Wagner C.J. and Rowe J.).

•         False allegations of sexual assault based on ulterior motives are more common than false allegations of other offences (Seaboyer, at p. 669, per L’Heureux-Dubé J., dissenting in part; R. v. Osolin, [1993] 4 S.C.R. 595, at p. 625, per L’Heureux-Dubé J., dissenting; R. v. A.G., 2000 SCC 17, [2000] 1 S.C.R. 439, at para. 3, per L’Heureux-Dubé J., concurring).

•         Real victims of sexual assault should have visible physical injuries (Seaboyer, at pp. 650 and 660, per L’Heureux-Dubé J., dissenting in part; R. v. McCraw, [1991] 3 S.C.R. 72, at pp. 83-84, per Cory J. for the Court).

•         A complainant who said “no” did not necessarily mean “no”, and may have meant “yes” (Seaboyer, at p. 659, per L’Heureux-Dubé J., dissenting in part; R. v. Esau, [1997] 2 S.C.R. 777, at para. 82, per McLachlin J. (as she then was), dissenting; R. v. Ewanchuk, [1999] 1 S.C.R. 330, at paras. 87 and 89, per L’Heureux-Dubé J., concurring; R. v. Cinous, 2002 SCC 29, [2002] 2 S.C.R. 3, at para. 167, per Arbour J., dissenting; R. v. Kirkpatrick, 2022 SCC 33, at para. 54, per Martin J. for the majority; R. v. Goldfinch, 2019 SCC 38, [2019] 3 S.C.R. 3, at paras. 44 and 74, per Karakatsanis J. for the majority).

•         If a complainant remained passive or failed to resist the accused’s advances, either physically or verbally by saying “no”, she must have consented — a myth that has historically distorted the definition of consent and rendered rape “the only crime that has required the victim to resist physically in order to establish nonconsent” (Ewanchuk, at paras. 93, 97 and 99, per L’Heureux-Dubé J., concurring, quoting S. Estrich, “Rape” (1986), 95 Yale L.J. 1087, at p. 1090; see also para. 103, per McLachlin J., concurring, and para. 51, per Major J. for the majority; see further R. v. M. (M.L.), [1994] 2 S.C.R. 3, at p. 4, per Sopinka J. for the Court; R. v. Find, 2001 SCC 32, [2001] 1 S.C.R. 863, at para. 101, per McLachlin C.J. for the Court; Cinous, at para. 167, per Arbour J., dissenting; R. v. Barton, 2019 SCC 33, [2019] 2 S.C.R. 579, at paras. 98, 105, 107, 109 and 118, per Moldaver J. for the majority; Friesen, at para. 151, per Wagner C.J. and Rowe J. for the Court).

•         A sexually active woman (1) is more likely to have consented to the sexual activity that formed the subject matter of the charge, and (2) is less worthy of belief — otherwise known as the “twin myths”, which allowed for regular canvassing of the complainant’s prior sexual history at trial, regardless of relevance, thereby shifting the inquiry away from the alleged conduct of the accused and towards the perceived moral worth of the complainant (Seaboyer; Ewanchuk).

[37]                         Myths and stereotypes about sexual assault complainants capture widely held ideas and beliefs that are not empirically true — such as the now-discredited notions that sexual offences are usually committed by strangers to the victim or that false allegations for such crimes are more likely than for other offences. Myths, in particular, convey traditional stories and worldviews about what, in the eyes of some, constitutes “real” sexual violence and what does not. Some myths involve the wholesale discrediting of women’s truthfulness and reliability, while others conceptualize an idealized victim and her features and actions before, during, and after an assault. Historically, all such myths and stereotypes were reflected in evidentiary rules that only governed the testimony of sexual assault complainants and invariably worked to demean and diminish their status in court.

[38]                         Overall, this legal backdrop reflected a time in which less was known about the prevalence of sexual violence and its lifelong harms. Eventually, Parliament, the courts, academics, and the public came to understand that the previous legal rules and the inaccurate, outdated, and inequitable social attitudes they represented impeded the equal treatment of sexual assault complainants and, hence, the overall fairness of trials.

[39]                         New Criminal Code provisions were drafted replacing the former offence of rape with a new offence of sexual assault, and the marital rape exemption was repealed — reflecting the reality that while women continue to make up the vast majority of sexual assault complainants, the offence of sexual assault can be perpetrated by and against people of all genders. Consent was expressly defined in s. 273.1(1)  of the Code  as the voluntary agreement of the complainant to engage in the sexual activity in question, and ss. 273.1(2) and 273.2 clearly circumscribed when no consent is obtained and the scope of mistaken belief in consent. The corroboration requirement was abolished and the doctrine of recent complaint was abrogated, pursuant to ss. 274  and 275  of the Code , respectively. Finally, ss. 276 and 278.1 to 278.91 now govern the admissibility of a complainant’s prior sexual activity and the production and admission of records containing the complainant’s highly sensitive personal information that is either held by a third party or comes into the hands of the accused.

[40]                         The significant legislative changes in this area of law were made with a view to protecting the rights of women and children given their particular vulnerability to sexual violence. In the preamble to the Act that amended s. 276  of the Code  post-Seaboyer and reformed the definition of consent to sexual activity, Parliament explicitly focused on combatting “the prevalence of sexual assault against women and children” and expressed a particular intention to “ensure the full protection of the rights guaranteed under sections 7  and 15  of the Canadian Charter of Rights and Freedoms ” (An Act to amend the Criminal Code (sexual assault), S.C. 1992, c. 38). Parliament also expressed an intention to “encourage the reporting of incidents of sexual violence” and emphasized that the admission of the complainant’s sexual history into evidence is “inherently prejudicial” and “should be subject to particular scrutiny” (ibid.). In amending ss. 278.1  to 278.91  of the Code , Parliament sought to protect complainants’ privacy interests in highly sensitive personal information and, in turn, preserve their dignity. Amendments such as these appropriately balance respect for complainants’ dignity, privacy, and equality and the fundamental right of the accused to the presumption of innocence and a fair trial and they have been endorsed as constitutionally compliant by this Court (see Mills; Darrach; J.J.).

[41]                         In turn, several of the myths and stereotypes outlined above have now been jurisprudentially condemned as errors of law. For example, it is prohibited for a trial judge to rely on notions such as: delay in a complainant’s disclosure of a sexual assault, alone, undermines the credibility of the disclosure (D.D.; R. v. Lacombe, 2019 ONCA 938, 383 C.C.C. (3d) 114); a complainant’s “failure” to dress modestly indicates that she is more likely to have consented (Ewanchuk, at para. 103); a complainant’s “failure” to resist or cry out is suggestive of consent (para. 93); the mere fact of a complainant having psychiatric or therapeutic consultations is relevant to their credibility or reliability (s. 278.3(4)  of the Code ); or a complainant associating with or not avoiding the accused after the alleged sexual assault suggests that there was consent and that no assault occurred (A. (A.B.), at paras. 6-12).

[42]                         The legislative and jurisprudential treatment of these issues reflects a collective understanding that courts should strive to eradicate myths and stereotypes from their decisions because they threaten the rights of complainants and undermine the truth-seeking function of trials. Today, sexual offences remain underreported and continue to occur mostly against women and children. There remains a “need to affirm the principles of equality and human dignity in . . . criminal law by addressing the problem of myths and stereotypes about complainants in sexual assault cases” (A.G., at para. 1, per L’Heureux-Dubé J., concurring; see R. v. O’Connor, [1995] 4 S.C.R. 411).

[43]                         This Court has repeatedly held that “myths and stereotypes have no place in a rational and just system of law, as they jeopardize the courts’ truth-finding function” (A.G., at para. 2). A trial is a truth-seeking process, and reliance on myths and stereotypes distorts the truth. In Mills, this Court explained that myths and stereotypes about sexual assault victims hamper the search for truth and impose “harsh and irrelevant burdens on complainants in prosecutions of sexual offences” (para. 119). While the accused’s constitutional rights must remain at the forefront of any criminal trial, this Court has also acknowledged that measures can be taken to avoid reliance on myths and stereotypes without compromising those rights. Myths and stereotypes in fact undermine a fair trial — meaning a trial that is fair not only to the accused, but to the complainant and the public (see J.J., at paras. 1-2).

[44]                         All of this history puts into perspective the distinct reasons why relying on myths and stereotypes to discredit sexual assault complainants amounts to an error of law — as opposed to being an ordinary factual finding reviewable for palpable and overriding error. The very reason this error of law emerged was to prevent the accused from discrediting complainants’ testimony on unwarranted, discriminatory grounds, and accordingly to correct the particular advantage accused persons historically had in sexual assault cases as compared to accused persons in other cases: myths and stereotypes are no longer meant to play any role in mounting a defence. The judicial and legislative developments designed to eradicate the categorical discounting of women as witnesses do not create any special benefits in law for complainants in sexual assault cases. They simply remove discriminatory barriers, establish a level testimonial field between sexual assault complainants and complainants in other cases, and ensure the truth-seeking function of the trial is not distorted. 

(b)          The Proposed Rule Should Not Be Adopted as a Corollary to the Prohibition Against Myths and Stereotypes

[45]                         In sum, the prohibition against myths and stereotypes about sexual assault complainants carries with it a discrete history, purpose, and character. It was designed with the specific goal of protecting complainants against prejudicial or discriminatory reasoning in criminal trials. The proposed rule against ungrounded common-sense assumptions should be rejected because it enlarges that particular context beyond its proper scope.

(i)            The Proposed Rule Disregards the Distinct Character of Myths and Stereotypes

[46]                         The types of assumptions captured by the proposed rule do not share the distinct discriminatory character of myths and stereotypes relating to sexual assault complainants, which form a unique category of legal error created to be explicitly remedial in nature. As noted in Find, this Court has been willing to accept the prevailing existence of myths and stereotypes about sexual assault complainants based on “overwhelming evidence” from the social science literature (para. 101). There is no comparable legislative recognition of a pattern of improper reliance upon a specific, circumscribed set of generalizations to incorrectly assess the testimony of accused persons in sexual assault cases. Myths and stereotypes about sexual assault complainants are governed by a limited set of legal principles that emerged to redress pervasive systemic prejudice, which was embedded in law and directed against a distinct category of witness — they cannot be coherently analogized to support a rule prohibiting any generalizations about any witness, including the accused.

[47]                         Unlike the proposed rule, the authorities that target sexual assault myths and stereotypes remove from the law specific erroneous assumptions and notions that formerly distorted judicial reasoning and fact-finding on a widespread basis. These myths and stereotypes unfairly undermine the perceived credibility of a specific category of witness before the evidence has even been heard, effectively imbuing complainants’ evidence with a prima facie judgment of frailty. They also distort the legal significance of consent. The legal principles that prohibit judges from relying on specific stereotypes when assessing the testimony of complainants are therefore in a separate category and are fundamentally different from a blanket prohibition on generalizations when assessing testimony at large. Conversely, the proposed rule does not relate to specific, identified, erroneous generalizations about a specific category of witness, many of which were first imposed by law, nor does it protect elements of an offence from taking on a distorted meaning. The rule instead casts a vast net — effectively transforming all types of factual generalizations, regardless of their nature, into errors of law. In doing so, the rule impermissibly expands the bounds of existing errors of law to any and all factual findings that can be said to amount to assumptions about human behaviour.

[48]                         As noted above, there is significant overlap between the rule against common-sense assumptions and the related rule against stereotypical reasoning articulated in J.C.: both rules bar trial judges from using “stereotypes” in their assessments of credibility. However, in J.C. and as adopted by the British Columbia Court of Appeal, the term “stereotype” not only demarcated an error of law, but was used in a manner that effectively encompassed all generalizations about human behaviour. With respect, this is not the correct approach: just because something is a generalization does not necessarily mean it amounts to a stereotype that can be understood as an error of law.

[49]                         Generally speaking, all ungrounded assumptions about human behaviour, including stereotypes, share two characteristics. First, they take a general proposition and apply it to a specific individual, foregoing any assessment of that person’s unique characteristics or circumstances. Second, that general proposition is inaccurate or untrue, either in all cases or as applied to that specific individual (see, e.g., R. v. G. (C.M.), 2016 ABQB 368, 41 Alta. L.R. (6th) 374, at para. 60). However, stereotyping goes one step further and connotes a particular legal meaning that merely generalizing does not: specifically, a meaning rooted in discrimination and inequality of treatment.

[50]                         The discriminatory character of stereotypes is made plain from our jurisprudence’s understanding of stereotypes about sexual assault complainants. Reliance on such stereotypes was recognized as an error of law for the very purpose of eliminating discrimination against women and promoting their dignity and equality within the justice system. For example, the requirement that a woman raise a “hue and cry” was based on the now-discredited assumption that, because rape was the worst thing that could happen to a woman, any credible victim would immediately disclose what happened to the first person she encountered. Accordingly, it is now an error of law to draw an adverse inference from the mere fact that the complainant did not report the assault immediately (D.D., at para. 65). The historical requirement that a complainant’s testimony be corroborated, now abolished, was based on the discriminatory assumption that the testimony of a woman was not, in law, equal to that of a man. The twin myths, now prohibited by s. 276  of the Code , were again based on the discriminatory notion that women who had sex were less trustworthy and did not deserve equal respect and treatment under the law (Seaboyer). The negative stereotypes about sexual assault complainants were recognized as errors of law because they often originated in law, took statutory form, and were founded on the misinterpretation of the now-clear definition of consent as the voluntary, communicated agreement to the sexual activity in question — a misinterpretation that undermined the core principle of equal justice for all.

[51]                         The problem with the proposed rule against ungrounded common-sense assumptions is that it fails to appreciate this crucial dimension. It instead lumps together the sorts of pernicious, discriminatory stereotypes that both the courts and Parliament have worked to condemn and correct with more benign generalizations that, while they may be factually wrong, have nothing to do with inequality of treatment.

[52]                         In J.C., for example, the Ontario Court of Appeal found that the trial judge’s characterization of the accused’s evidence as mechanical, rehearsed, and politically correct relied on “stereotype” and therefore amounted to an error of law. In Roth, although the British Columbia Court of Appeal rejected the submission that the trial judge had relied on stereotypes about male behaviour, it nevertheless found that the trial judge’s improper generalization about the accused being a power lifter amounted to an error of law. Regardless of whether these factual assumptions were wrong or right, there is no question they are fundamentally of a different character than inequality-based stereotypes. A witness being discredited because they testified in what the trial judge considered to be an implausible manner, or because of their fitness habits, is simply not the same thing as a complainant being discredited because of her sexual history and therefore “unchaste” character, because she did not resist her assailant, or because she did not report the assault right away. This is in no way meant to condone reliance on improper generalizations in credibility assessments. Rather, it illustrates that ordinary factual generalizations — which are of a different character than those relating to myths and stereotypes about sexual assault complainants, and do not disclose other recognized errors of law — should be evaluated for palpable and overriding error. 

[53]                         The legislative shift towards a more gender-neutral understanding of sexual violence compels courts to also be sensitive to the fact that complainants in sexual assault cases come from all walks of life. By eliminating the inherently gendered crime of rape, Parliament signalled that sexual assault is no longer meant to be limited to the male accused, female complainant paradigm. Stereotypes about complainants that formerly targeted women alone now have the potential to compromise the testimonial assessment of all types of complainants — for example, although it remains the case that the majority of sexual assault complainants are women, it is an error of law to rely on the doctrine of immediate complaint or on stereotypes about a complainant’s past sexual activity regardless of the complainant’s gender. The wholesale discrediting of complainants as a particular category of witness, at its core, is also rooted in inequality of treatment, as this Court recognized in O’Connor (at para. 123):

. . . a legal system which devalues the evidence of complainants to sexual assault by de facto presuming their uncreditworthiness would raise similar concerns. It would not reflect, far less promote, “a society in which all are secure in the knowledge that they are recognized at law as human beings equally deserving of concern, respect and consideration” (Andrews v. Law Society of British Columbia, [1989] 1 S.C.R. 143, at p. 171). [Emphasis deleted.]

 

(See also Seaboyer, at p. 654.)

[54]                         In turn, all sexual assault cases have the potential to engage a broad range of stereotypes beyond those about female complainants. In Barton, for example, this Court recognized specific myths and stereotypes that apply to Indigenous women, which may in turn hinder proper assessments of their credibility and reliability. Courts have also moved away from treating the testimony of very young complainants with inherent suspicion, recognizing that past attitudes did not account for the inherent vulnerability of children and the particular problems they may face when giving evidence (see R. v. F. (W.J.), [1999] 3 S.C.R. 569). I would also stress at this juncture that reliance on stereotypes, being rooted in inequality of treatment, is certainly not just a problem for sexual assault complainants alone. Stereotypical reasoning based in the sort of inequality of treatment at the heart of myths and stereotypes against sexual assault complainants has the potential to affect the testimony of all witnesses in all trials. It remains open to all parties in future cases, including the accused, to argue that trial judges have employed stereotypes rooted in other, analogous forms of inequality of treatment, and therefore erred in law. 

[55]                         In addition to its unsound foundations, the proposed rule interacts particularly poorly with sexual offence cases given the carefully crafted rules that now govern the assessment of a complainant’s testimony. The Court of Appeal in Mr. Tsang’s case acknowledged the reality that for certain common-sense assumptions about sexual activity to be “grounded” in the evidence, evidence of a “complainant’s own predilections” may be required (para. 19). In that case, to “ground” the trial judge’s finding that it was unbelievable the complainant would ask for rough sex, the complainant would first have to be subjected to deeply intrusive questioning about her sexual preferences or past practices. In Mr. Kruk’s case, to “ground” the complainant’s testimony that she was not mistaken about the physical sensation of penile-vaginal penetration in the evidence, the complainant would need to testify about previously feeling that sensation.

[56]                         The proposed rule thereby drastically expands the scope of permissible questioning into a complainant’s sexual history, effectively requiring both parties to apply, pursuant to Seaboyer (at p. 619) or s. 276  of the Code , to adduce “other sexual activity” evidence that may not otherwise be relevant to the main issues in the case or permitted to make its way into the trial at all. It opens a back door to prohibited twin-myth reasoning insofar as, in the process of “grounding” the trial judge’s findings about her testimony, evidence of the complainant’s sexual history may be treated as logically probative of her credibility or consent. Although framed in terms of ensuring “equal” treatment for accused in these cases, this approach in fact risks resurrecting the very prejudice against sexual assault complainants that the law on myths and stereotypes was designed to eliminate.

[57]                         In sum, the proposed rule against ungrounded common-sense assumptions cannot be understood as a logical extension of legal rules against stereotyping. To the extent it conflates stereotyping with all assumptions about human behaviour, it runs off course. The concept of a stereotype is not closed and no doubt will continue to evolve in future cases: the closer an error is to the types of myths and stereotypes pertaining to sexual assault complainants that have been recognized in the jurisprudence, the more likely it is that it will amount to an error of law. However, all other mere assumptions drawn in the course of credibility and reliability assessments, like other findings of fact, must remain reviewable for palpable and overriding error.

(ii)          The Accused’s Rights Remain Protected

[58]                          The respondents argue that the proposed rule against ungrounded common-sense assumptions should be adopted in sexual assault cases out of fairness to the accused. This argument, too, must be rejected. There are crucial legal protections explicitly designed to ensure fairness to the accused that find their source in their own robust body of law. It is essential that these legal protections be taken seriously and given their full effect to ensure the accused’s constitutional rights remain safeguarded. However, the idea that an identical rule mirroring the treatment of myths and stereotypes between complainants and accused is necessary to achieve that goal is misguided.

[59]                          The overarching principle of the presumption of innocence, enshrined in s. 11(d)  of the Charter , and the correlative principle of the Crown’s burden of proof, must always govern the fact-finding process. The presumption of innocence — a “hallowed principle lying at the very heart of criminal law” (R. v. Oakes, [1986] 1 S.C.R. 103, at p. 119) — requires the Crown to bear the onus of proving all essential elements of the offence charged, beyond a reasonable doubt, before a conviction may be entered (Osolin). Closely related to the presumption of innocence is the accused’s right to silence as enshrined in s. 11(c)  of the Charter , which safeguards human dignity and privacy against processes or reasoning that would compel an accused person to incriminate themselves with their own words (R. v. Noble, [1997] 1 S.C.R. 874, at paras. 69-78).

[60]                          Various protections relating to the assessment and weighing of evidence flow from the presumption of innocence and the right to silence. Notably, an accused’s silence at trial may not be treated as evidence of guilt, as such reasoning would violate both principles (Noble, at para. 72). Likewise, it is improper to discount the credibility of accused persons on the basis that, because they face criminal penalties, they will say anything to protect themselves. Though considering the possibility that the accused may have a motive to lie will not necessarily offend this rule, courts should be wary of going further and drawing the “impermissible assumption” that they will do so in all cases (R. v. Laboucan, 2010 SCC 12, [2010] 1 S.C.R. 397, at para. 14). Such reasoning is premised on a supposition of guilt and therefore offends the presumption of innocence (para. 12).

[61]                          The presumption of innocence also restricts how credibility is assessed in cases of conflicting testimony between defence and Crown witnesses. The analysis of testimony must never be treated as a contest of credibility, and triers of fact need not accept the defence’s evidence or version of events in order to acquit (R. v. Van, 2009 SCC 22, [2009] 1 S.C.R. 716, at para. 23; R. v. W. (D.), [1991] 1 S.C.R. 742, at p. 757). The burden never shifts to the accused to establish their own innocence, and the onus always lies with the Crown to prove every essential element (R. v. J.H.S., 2008 SCC 30, [2008] 2 S.C.R. 152, at para. 13).

[62]                          Reasonable doubt applies to credibility assessments such that if the evidence the Crown adduced does not rise to the level required of a criminal conviction, an accused cannot be found guilty simply because they are disbelieved (see W. (D.)). Some elements of the totality of the evidence may give rise to a reasonable doubt, even where much — or all — of the accused’s evidence is disbelieved. Any aspect of the accepted evidence, or the absence of evidence, may ground a reasonable doubt. Moreover, where the trier of fact does not know whether to believe the accused’s testimony, or does not know who to believe, the accused is entitled to an acquittal (J.H.S., at paras. 9-13; R. v. H. (C.W.) (1991), 68 C.C.C. (3d) 146 (B.C.C.A.); R. v. S. (W.D.), [1994] 3 S.C.R. 521, at p. 533; R. v. Avetysan, 2000 SCC 56, [2000] 2 S.C.R. 745, at para. 19). Finally, where the Crown relies on circumstantial evidence to establish guilt, the trier of fact may only convict if guilt is the only reasonable inference from the evidence (R. v. Villaroman, 2016 SCC 33, [2016] 1 S.C.R. 1000, at para. 30).

[63]                          Protections like these ensure fairness to the accused and must guide trial judges in assessing testimony. They impose specific prohibitions regarding identifiable forms of improper reasoning. In so doing, they safeguard the accused’s fundamental Charter  rights. Where full effect is properly accorded to them, further restrictions on how trial judges may assess the credibility and reliability of the accused and the complainant that go beyond existing errors of law and the standard of palpable and overriding error are rendered unnecessary. Even where recognized errors of law pertaining to the accused’s rights are not engaged, where the accused chooses to testify or otherwise lead evidence, appellate courts must ensure that trial judges have thoroughly and properly assessed the accused’s account of events. Reliance on illogical, untrue, or otherwise improper assumptions in assessing the evidence of an accused is wrong and remains reversible on appeal if it amounts to palpable and overriding error. It is just that such ordinary factual assumptions cannot be recognized, wholesale, as stand-alone errors of law. 

[64]                         It must also be emphasized that the concept of myths and stereotypes concerning sexual assault complainants is not unbounded. It has produced a circumscribed set of legal rules that require careful application, close attention to context, and a nuanced understanding of the purpose for which any given piece of evidence is tendered. Some scholars have suggested that the law of myths and stereotypes is presently being overused in contexts where it is inapplicable, or applied without rigour (see, e.g., L. Dufraimont, “Current Complications in the Law on Myths and Stereotypes” (2021), 99 Can. Bar Rev. 536). If this problem exists, the appropriate solution is not to extend parallel errors of law that apply to accused persons as well as complainants. Rather, mindful that myths and stereotypes against sexual assault complainants give rise to an error of law, courts must ensure these myths and stereotypes are not extended beyond their permissible scope. 

[65]                         For example, just because the evidence happens to align with a myth or stereotype does not necessarily mean that any inferences that can be drawn from that evidence will be prejudicial. While it is a myth that women regularly fabricate allegations of sexual assault, it is not an error to consider whether the circumstances of a particular case support the existence of a motive to fabricate (see, e.g., R. v. Esquivel-Benitez, 2020 ONCA 160, 61 C.R. (7th) 326, at paras. 9-15) — indeed, where the defence adduces evidence on this point, a trial judge is obliged to consider it to give full effect to the presumption of innocence, and a failure to do so constitutes reversible error. Furthermore, while s. 276  of the Code  prohibits the use of prior sexual history to support either twin myth, it clearly does not prohibit such evidence point-blank and for all purposes. So long as it is properly screened by s. 276, such evidence may be used, for instance, to resolve inconsistencies between the complainant and the accused’s testimony as to their relationship.

[66]                         Our existing legal framework is sufficient to ensure that the accused’s rights remain protected in sexual assault cases. It is vital that the accused’s Charter  rights be carefully respected, that any evidence the accused gives be properly assessed, and that the concept of myths and stereotypes remain appropriately constrained to its proper scope. However, there is no need to adopt the proposed rule against ungrounded common-sense assumptions in order to correct any putative unfairness.

(iii)        The Difference Between Factual and Legal Speculation

[67]                         The conceptual slippage between stereotyping and generalizing in the jurisprudence of the lower courts illustrates the importance of precisely defining errors of law and keeping them distinct from other factual findings, even erroneous ones. In this vein, brief comment is also warranted in response to the respondent Mr. Tsang’s argument that the proposed rule against ungrounded common-sense assumptions is a logical offshoot of the error of law of “speculation”. With respect, this submission confuses the colloquial meaning of speculation — drawing conclusions without basis in fact — with the more precise form of speculation as an error of law.

[68]                         Speculation as an error of law arises where a trial judge has found that certain evidence “creates a reasonable doubt as to the guilt of the accused, when, on a proper view of the law, that evidence is not capable of creating any doubt as to his guilt” (Wild v. The Queen, [1971] S.C.R. 101, at p. 111). In other words, it is an error of law to fail to distinguish between a rational conclusion as to reasonable doubt based on evidence, and an unsupported conclusion based on conjecture (see Wild; Rousseau v. The Queen, [1985] 2 S.C.R. 38; R. v. B. (G.), [1990] 2 S.C.R. 57; R. v. Clark, 2015 BCCA 488, 407 D.L.R. (4th) 610, at para. 43, aff’d 2017 SCC 3, [2017] 1 S.C.R. 86). The term “speculation” as invoked in the courts below does not refer to this legal error, but rather is used, as it is in common parlance, to signify drawing any inferences that are allegedly unsupported by evidence.^[2] To the extent that, in recognizing the proposed rule, the courts below rely on the error of law of speculation — which most often arises where trial judges err in acquitting the accused based on speculative alternative explanations — they are stretching the case law beyond what it logically bears.

(2)          The Proposed Rule Runs Contrary to Established Standards of Review and Is Counterproductive to Proper Testimonial Assessment

[69]                          Although the proposed rule against ungrounded common-sense assumptions has often been invoked in the sexual assault context, its applications extend to all criminal cases. I approach these reasons primarily in the context of sexual assault given that this is the offence immediately before the Court, but nothing in them should be taken as an endorsement of the proposed rule in any other case. Part of the justification for the rule’s rejection in the context of these appeals is rooted in the body of law relating to sexual assault myths and stereotypes. However, cognizant that the rule has also been invoked in other contexts,^[3] its fundamental weaknesses extend to all criminal appeals.

[70]                          With that in mind, another reason why this proposed rule should not be adopted is that it is profoundly incompatible with the established law associated with credibility and reliability findings. The rule is not a cautious or incremental extension of existing principles — it is instead an unwarranted departure from how appellate courts typically treat trial judges’ credibility findings. By exposing trial judges’ findings to overly invasive appellate review, it creates a serious barrier to meaningful testimonial assessment, creating mischief that will extend across the entire criminal law.

(a)           The Role of Common Sense in Evaluating a Witness’s Testimony

[71]                          First, the proposed rule is incompatible with the often inextricable role common-sense assumptions play in credibility and reliability assessments. Testimonial assessment is largely based on inductive reasoning and the particular circumstances of the case: it requires the trier of fact to make assessments based on probable interpretations of the evidence (R. v. Calnen, 2019 SCC 6, [2019] 1 S.C.R. 301, at para. 111; R. v. Munoz (2006), 86 O.R. (3d) 134 (S.C.J.), at para. 23). Testimonial assessment therefore necessarily depends on the life experience a trial judge brings to their task, which, in turn, informs the common-sense inferences they draw from what they see before them.

[72]                         It is widely recognized that testimonial assessment requires triers of fact to rely on common-sense assumptions about the evidence. In R. v. Delmas, 2020 ABCA 152, 452 D.L.R. (4th) 375, at para. 31, aff’d 2020 SCC 39, [2020] 3 S.C.R. 780, the Alberta Court of Appeal observed that triers of fact may rely on reason and common sense, life experience, and logic in assessing credibility. In R. v. R.R., 2018 ABCA 287, 366 C.C.C. (3d) 293, the same court held that triers of fact “must invariably fall back on their common sense, and their acquired knowledge about human behaviour in assessing the credibility and reliability of witnesses” (para. 6). Finally, in R. v. S. (R.D.), [1997] 3 S.C.R. 484, this Court considered that the life experience of trial judges — though of course not a substitute for evidence, and subject to appropriately circumscribed limits — “is an important ingredient in the ability to understand human behaviour, to weigh the evidence, and to determine credibility”, and assists with a “myriad of decisions arising during the course of most trials” (para. 13). Reasoning about how people generally tend to behave, and how things tend to happen, is not only permissible, it is often a necessary component of a complete testimonial assessment.

[73]                         In turn, common-sense assumptions necessarily underlie all credibility and reliability assessments. Credibility can only be assessed against a general understanding of “the way things can and do happen”; it is by applying common sense and generalizing based on their accumulated knowledge about human behaviour that trial judges assess whether a narrative is plausible or “inherently improbable” (R. v. Kiss, 2018 ONCA 184, at para. 31 (CanLII); R. v. Adebogun, 2021 SKCA 136, [2022] 1 W.W.R. 187, at para. 24; R. v. Kontzamanis, 2011 BCCA 184, at para. 38 (CanLII)). Common sense underpins well-established principles guiding credibility assessment — including the now-universal idea that witnesses who are inconsistent are less likely to be telling the truth — and assists in assessing the scope and impact of particular inconsistencies. Reliability also requires reference to common-sense assumptions about how witnesses perceive, remember, and relay information, invoking generalizations about how individuals tend to present information that they are remembering accurately and completely, as opposed to matters about which they are unsure or mistaken. A trial judge may, for example, infer that a witness was credible yet unreliable because they appeared sincere but displayed indicia that tend to suggest an unclear or uncertain memory (e.g., equivocation, phrases such as “hmm . . . let me see”, long pauses, or failure to provide much detail).

[74]                         Even the proponents of the rule against ungrounded common-sense assumptions accept that common sense is necessary and, to some extent, inevitable, to the task of testimonial assessment. The Court of Appeal in Mr. Kruk’s case acknowledged that in “working through the minefield of legal and evidentiary issues, trial judges apply their common sense to the evidence to reach sound verdicts” (para. 2), and indeed, “[r]elying on their life experience to assess the credibility of witnesses is a daily and appropriate exercise for trial judges” (para. 41). In J.C., the court held that there is no bar on relying on common sense or human experience to identify inferences arising from the evidence — otherwise, circumstantial evidence, which depends on bridging gaps between the evidence and the inference drawn using human experience, would not be admissible at all. The court also correctly observed that there is no bar on using human experience to draw inferences from evidence. Otherwise, many other well-established principles of evidence — such as the idea that fleeing the scene or destroying evidence after a criminal offence is generally proof of concealment of guilt — would also be put at risk.

[75]                         By prohibiting ungrounded common-sense assumptions, the proposed rule interferes with the necessary recourse to common sense as a part of testimonial analysis. Trial judges are uniquely tasked with assessing the testimony they hear and interpreting the range of possible inferences arising from the evidence. They must be able to rely not only on their judicial experience as fact-finders, but also on their common sense and the generalized expectations it generates about human behaviour. Trial judges will naturally rely on “ungrounded” assumptions about human behaviour in their testimonial assessments and thereby draw on factors that lie outside the immediate record. The judicial function entitles them to do so without requiring extrinsic evidence to support each and every one of their conclusions.

[76]                         The proposed rule’s rationale is belied by a contradiction inherent in its own logic. It prohibits relying on common sense to introduce new considerations not arising from evidence — while simultaneously acknowledging that common sense can be used as an interpretive aid, which necessarily involves importing considerations arising not from the evidence itself but from a judge’s accumulated life experience. It is effectively impossible to draw a clear boundary between using human experience to interpret evidence or draw inferences (which is permissible under the rule) and introducing new considerations into the evidence (which is not).

[77]                         According to the rule as described by the majority of the Court of Appeal below in Mr. Tsang’s case, impermissible reasoning is: “affected by implicit, unsupported assumptions about ‘normal behaviour’” (para. 53); “unsubstantiated, untethered to the evidence and a prejudicial stereotype” (para. 65); “assumptions with respect to human behaviour” (para. 73); “generalizations about normative behaviour [that] did not rest on the evidence” (para. 74); “unfounded” assumptions or “generalizations” (paras. 84 and 112); and “unsupported inferences” (para. 115). On the other hand, permissible reasoning includes: “. . . a conclusion about what this complainant is likely to have done in these circumstances” (para. 39 (emphasis deleted)); and “the class of inferences relating to behaviour that may fairly be drawn” and “not used to draw an inference that is ‘unfair and inaccurate’” (para. 40). Curiously, the Court of Appeal in Mr. Tsang’s case considered it permissible to infer that people who have recently met are unlikely to want to share drinking glasses, because such an inference is “based on common sense and [is] not prejudicial” (para. 40). Yet this inference is undeniably a generalization about human behaviour that was not grounded in any evidence.

[78]                         Given that a trial judge is inevitably bound to rely on a common-sense assumption at some point when assessing a witness’s testimony, the ill-defined requirement that such assumptions must be “grounded in evidence” would also compel counsel in criminal cases to lead direct evidence to establish a wide range of notions that are generally true. To return to the example from Mr. Tsang’s case, to properly be able to find that people who have just met are unlikely to share drinking glasses at a bar — putting aside whether this is actually the case — the court would require direct evidence, presumably from the witnesses themselves. This evidence might include how closely each witness typically interacts with strangers in bars, their personal hygiene habits, and any inclinations or aversions they may have to sharing drinks with people they do not know, as well as presumably the underlying reasons for all of these propositions. Such matters are mundane, take up inordinate amounts of trial time, and are often of minimal, if any, relevance to the alleged offence. Particularly in sexual assault cases, which are already fraught with complexities, the incredible complications this type of evidentiary obligation would occasion cannot be understated.

[79]                         In sum, the proposed rule is fundamentally unfaithful to the necessary and proper use of common sense when assessing the testimony of witnesses. Worse, the rule also fails to establish any discernable boundary between the permissible and impermissible uses of common-sense assumptions. There is no coherent method to determine what assumptions are sufficiently uncontroversial to be “grounded in the evidence”, nor how much evidence is required to “ground” them. Instead, as explained further below, the rule seems to reduce this question to what a particular reviewing court deems to be fair, accurate, or uncontentious. The rule thereby invites appellate courts to substitute their opinions about what generalizations are appropriate or instructive in any given circumstance for those of trial judges, improperly transforming their “strong opposition to [a] trial judge’s factual inferences . . . into supposed legal errors” (R. v. George, 2017 SCC 38, [2017] 1 S.C.R. 1021, at para. 17; A.F. (Tsang), at para. 74). This dynamic creates uncertainty and unfairness on appeal.

(b)          Established Standards of Review for Credibility and Reliability Assessments

[80]                          The proposed rule runs contrary to well-established standards of review and would unduly increase the scope of appellate intervention into the credibility and reliability assessments of trial judges. Adopting the rule would compromise the delicate task trial judges undertake when evaluating the evidence of a witness by subjecting their reasons to unjustifiably invasive appellate scrutiny.

[81]                         Assessments of credibility and reliability can be the most important judicial determinations in a criminal trial. They are certainly among the most difficult. This is especially so in sexual assault cases, which often involve acts that allegedly occurred in private and hinge on the contradictory testimony of two witnesses. The trial judge, while remaining grounded in the totality of the evidence, is obliged to evaluate the testimony of each witness and to make determinations that are entirely personal and particular to that individual. Credibility and reliability assessments are also context-specific and multifactorial: they do not operate along fixed lines and are “more of an ‘art than a science’” (S. (R.D.), at para. 128; R. v. Gagnon, 2006 SCC 17, [2006] 1 S.C.R. 621).^[4] With respect to credibility in particular, while coherent reasons are crucial, it is often difficult for trial judges to precisely articulate the reasons why they believed or disbelieved a witness due to “the complex intermingling of impressions that emerge after watching and listening to witnesses and attempting to reconcile the various versions of events” (Gagnon, at para. 20; see also R. v. R.E.M., 2008 SCC 51, [2008] 3 S.C.R. 3, at para. 28; R. v. G.F., 2021 SCC 20, [2021] 1 S.C.R. 801, at para. 81). The task is further complicated by the trial judge’s ability to accept some, all, or none of a witness’s testimony.

[82]                          The governing standard of review applicable to findings of credibility and reliability is well established: absent a recognized error of law, such findings are entitled to deference unless a palpable and overriding error can be shown (Gagnon, at para. 10, citing Schwartz v. Canada, [1996] 1 S.C.R. 254, at paras. 32-33; H.L. v. Canada (Attorney General), 2005 SCC 25, [2005] 1 S.C.R. 401, at para. 74). Credibility and reliability findings typically do not engage errors of law, as at their core they relate to the extent to which a judge has relied upon a particular factor and how closely that factor is tied to the evidence. Although such findings may be overturned on correctness if errors of law are disclosed, in most cases it is preferable to review them using the nuanced and holistic standard of palpable and overriding error — which defers to the conclusions of trial judges who have had direct exposure to the witnesses themselves.

[83]                         Trial judges have expertise in assessing and weighing the facts, and their decisions reflect a familiarity that only comes with having sat through the entire case. The reasons for the deference accorded to a trial judge’s factual and credibility findings include: (1) limiting the cost, number, and length of appeals; (2) promoting the autonomy and integrity of trial proceedings; and (3) recognizing the expertise and advantageous position of the trial judge (Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235, at paras. 12-18). In light of the practical difficulty of explaining the constellation of impressions that inform them, it is well-established that “particular deference” should be accorded to credibility findings (G.F., at para. 81). Appellate courts are comparatively ill-suited to credibility and reliability assessment, being restricted to reviewing written transcripts of testimony and often focussing narrowly, even telescopically, on particular issues as opposed to seeing the case and the evidence as a whole (Housen, at para. 14, citing R. D. Gibbens, “Appellate Review of Findings of Fact” (1991-92), 13 Advocates’ Q. 445, at p. 446).

[84]                         The unique nature of testimonial assessment also guides how reviewing courts approach their task on appeal. Appellate courts must be mindful of the acute practical difficulties trial judges face in articulating why a particular witness was believed or disbelieved, tasked as they are with interpreting the various impressions and inferences that arise from the evidence (Gagnon, at para. 20; see also R.E.M., at para. 28; G.F., at para. 81). An appellate court should examine a trial judge’s reasons as a whole and refrain from parsing their “individual linguistic components”, as such an invasive approach would “undermine the trial judge’s responsibility for weighing all of the evidence” (Gagnon, at para. 19; see also Housen, at para. 72; G.F., at para. 69; R. v. Chung, 2020 SCC 8, [2020] 1 S.C.R. 405, at para. 33; R.E.M., at paras. 35 and 54). The need to review the entire record and for a full, flexible and functional approach when scrutinizing the findings of a trial judge is tied to the nature of the decision-making process at trial: reasons for judgment “are not intended to be, and should not be read, as a verbalization of the entire process engaged in by the trial judge in reaching a verdict” (R. v. Morrissey (1995), 22 O.R. (3d) 514 (C.A.), at p. 525).

[85]                          The palpable and overriding error standard strikes the appropriate balance between deference to the factual findings of the trial judge and the need for meaningful review of criminal cases on appeal. Although this standard is duly deferential to the trial judge’s unique vantage point and expertise, even under this more deferential standard, appellate courts must determine whether the trial judge’s findings on credibility and reliability are “the product of an evidence-based and context-specific assessment” of the witness’s testimony (R. v. Pastro, 2021 BCCA 149, 71 C.R. (7th) 296, at para. 67). Trial judges must “clearly articulat[e]” the basis for their assessments and point to “a nexus to the facts of the case” as opposed to relying on “assumptions about expected responses or conduct” (Tanovich, at p. 92). Yet, the proposed rule against ungrounded common-sense assumptions undercuts the rationale for the palpable and overriding error standard by inviting appeal courts to examine the specific language of particular common-sense reasoning and scrutinize it on a standard of correctness. The ensuing appellate review exercise quickly becomes highly interventionist, cumbersome, and almost entirely unpredictable.

[86]                         Invoking the proposed rule, appellate courts have been invited to parse trial reasons, attack generic statements made in the course of credibility assessments, and frame any credibility findings based on human behaviour as impermissible stereotypes or common-sense assumptions untethered to evidence. In Perkins, Cepic, Roth, and J.C., the appeal courts quashed the convictions due to errors in the trial judges’ credibility assessments, concluding that the trial judges went beyond allowable common-sense inferences and engaged in reasoning not grounded in the evidence. In other cases, a review of the findings as a whole demonstrated that the trial judges had employed no unfounded assumptions and rather the credibility assessments had been conducted in relation to the parties and circumstances before the court. The jurisprudence in this area is variable, even volatile, and evinces the need for a more consistent approach to appellate review (see Perkins; Roth; Cepic; J.C.; see, contra, Pastro, at paras. 68-69; R. v. Greif, 2021 BCCA 187, at paras. 68-69 (CanLII); Adebogun; R. v. Al-Rawi, 2021 NSCA 86, 410 C.C.C. (3d) 385; R. v. K.B.W., 2022 SKCA 8; R. v. L.L., 2022 ONCA 50; R. v. Lapierre, 2022 NSCA 12; Kritik-Langer v. R., 2022 QCCA 657; R. v. Kavanagh, 2022 BCCA 225; R. v. D.B., 2022 SKCA 76, 415 C.C.C. (3d) 455; R. v. S.A., 2022 ONCA 642; R. v. S.M., 2023 ONCA 417).

[87]                         With respect, the Court of Appeal’s decision in Mr. Tsang’s case illustrates the microscopic form of appellate review this very Court has cautioned against. The Court of Appeal targeted specific word choices, approving certain formulations of the trial judge’s credibility findings while implying that even slightly different word choices would have been erroneous. For example, in holding that the trial judge did not rely on a prejudicial generalization in refusing to believe the complainant had physically expressed her interest in the accused on the dance floor, the court emphasized that the “trial judge did not conclude simply that [the idea] was ‘not capable of belief’” but that “it was not capable of belief in the circumstances” (para. 33 (emphasis in original)). Similarly, approving of the trial judge’s conclusion that it was unlikely the parties would have shared a drink because they were effectively strangers, the court held, “[i]mportantly . . ., the trial judge did not say [Mr. Tsang’s] evidence about sharing a drink was ‘unbelievable’, it was simply ‘unlikely’ to have occurred” (para. 40).

[88]                         The proposed rule also risks diverting the focus of the trial judge’s reasons on plausibility to issues of form over substance, creating a chilling effect against thorough and frank reasons. Under the rule, it would be permissible for a trial judge to find it implausible that x, y, or z would have occurred in the circumstances of the case — but it would be impermissible to plainly state the common-sense premise that underpins the finding of implausibility, for fear that an appellate court would deem that premise “ungrounded” in the evidence at trial. The unstated barometer underpinning plausibility is a generalized expectation about how events tend to unfold and how people tend to behave in particular situations, meaning that common-sense assumptions represent a necessary measuring stick against which to assess the plausibility of a narrative. Yet the proposed rule leads judges into a catch-22, as the judge’s reasons (read functionally and contextually) must provide both the “what” and the “why” (G.F., at paras. 68-70).

[89]                         This form of appellate review directly cuts against established principles and leads to arbitrariness in outcomes. It does not, in my respectful view, advance the interests of justice. To the extent trial judges are the ones who have heard the evidence, they are best placed to make the complex and multifaceted factual findings that culminate in fair and nuanced credibility assessments. Deference to trial judges’ assessments of that evidence and the words they choose to describe it is warranted.

[90]                          In general, the introduction of new errors of law has the potential to upset the established balance in relation to credibility and reliability findings. Review based on an error of law may invite a “yes-no” answer measured on a standard of correctness, which opens the door to undue scrutiny of matters properly before the trial judge. To some extent, the materiality inquiry associated with the rule against ungrounded common-sense assumptions — under which the appellant must show that such reasoning “mattered in arriving at the impugned factual finding” (J.C., at para. 100) — mitigates the categorical nature of an error of law. However, in my view, it remains preferable to assess whether an error has been made in the first place based on the palpable and overriding standard. An overriding error is necessarily material because it must be shown to have affected the trial judge’s decision — but it is important to emphasize that an overriding error affects not just an isolated finding of fact, which may or may not have played a role in reaching the outcome, but the trial judge’s decision as a whole. It is not enough for an appellant asserting palpable and overriding error to pull at leaves and branches and leave the tree standing; the entire tree must fall (South Yukon Forest Corp. v. Canada, 2012 FCA 165, 431 N.R. 286, at para. 46, cited in Benhaim v. St-Germain, 2016 SCC 48, [2016] 2 S.C.R. 352, at para. 38, and Salomon v. Matte-Thompson, 2019 SCC 14, [2019] 1 S.C.R. 729, at para. 116).

[91]                          Overall, the palpable and overriding error standard fosters an appropriately holistic approach to appellate review. As compared to the invasive method associated with the proposed error of law, palpable and overriding error is far better attuned to the deference rightly afforded to trial judges’ factual findings, including with respect to credibility and reliability findings. There is simply no need for this Court to endorse a departure from that established approach — let alone one so substantial — by recognizing a new error whose far-reaching repercussions would reverberate across the entire criminal law.

(3)          Summary

[92]                         For the reasons outlined above, the proposed rule against ungrounded common-sense assumptions should not be recognized as giving rise to an error of law. The rule is in no way analogous to the body of law protecting sexual assault complainants from myths and stereotypes, nor can it be justified as necessary to ensure fairness to the accused. It also treats any and all factual assumptions drawn in the course of testimonial assessments as errors of law and thereby represents an unjustified departure from well-established principles governing testimonial assessment and appellate standards of review.

[93]                         Without the rule in play, appellate courts are left to rely on the existing and well-established law on assessing a trial judge’s credibility or reliability assessments. For the utmost clarity, the applicable framework can be summarized as follows.

[94]                         First, where an appellant alleges that a trial judge erroneously relied on a “common-sense” assumption in their testimonial assessment, the reviewing court should first consider whether what is being impugned is, in fact, an assumption. Given the nature of how witnesses give evidence and the need to read the trial judge’s reasons as a whole, what might appear to be an assumption on its face may actually be a judge’s particular finding about the witness based on the evidence.

[95]                         Second, once satisfied that the trial judge did, in fact, rely on an assumption that is beyond the bounds of what common sense and the judicial function support, the reviewing court should identify the appropriate standard of review applicable to the impugned portion of the trial judge’s credibility or reliability assessment.

[96]                         The standard of review will be correctness if the error alleged is a recognized error of law. Nothing in these reasons should be taken to limit the scope of existing errors of law relating to testimonial assessments that this Court has previously approved. Such errors may include reliance on myths and stereotypes about sexual assault complainants, as well as any improper and incorrect assumptions about accused persons that run contrary to fundamental principles such as the right to silence and the presumption of innocence. Testimonial assessments may also become vulnerable to correctness review for reasonable apprehension of bias (S. (R.D.), at paras. 91-141), making a finding of fact for which there is no evidence (R. v. J.M.H., 2011 SCC 45, [2011] 3 S.C.R. 197, at para. 25; Schuldt v. The Queen, [1985] 2 S.C.R. 592, at p. 604), and improperly taking judicial notice (see, e.g., R. v. Poperechny, 2020 MBCA 81, 396 C.C.C. (3d) 478). As discussed, reliance on stereotypes other than myths and stereotypes about sexual assault complainants, but which are similarly rooted in inequality of treatment, may also amount to errors of law, and it remains open to all parties to argue as much in future cases. The list of errors of law is not closed — but the rule against ungrounded common-sense assumptions is not on it.

[97]                         Absent an error of law, the standard of review will be palpable and overriding error. The reviewing court must first determine whether the erroneous reliance on the assumption is palpable, in that it is “plainly seen”, “plainly identified”, or “obvious” (see Housen, at paras. 5-6; R. v. Clark, 2005 SCC 2, [2005] 1 S.C.R. 6, at para. 9; Benhaim, at para. 38, citing South Yukon Forest Corp., at para. 46). Palpable errors in this context will include, for example, where the assumption in question is obviously untrue on its face, or where it is untrue or inapplicable in light of the other accepted evidence or findings of fact. Although trial judges are clearly best placed to make factual findings and assess the accuracy of generalizations, appellate courts can balance the need for deference to those findings with employing their own common sense to determine whether the presumption was clearly illogical or unwarranted so as to make out a palpable error. Appellate courts are routinely tasked with, for example, considering whether based on “logic and human experience” a particular piece of evidence was relevant or whether an accused’s after-the-fact conduct was consistent with that of a guilty person (R. v. Corbett, [1988] 1 S.C.R. 670, at p. 715; see R. v. White, 2011 SCC 13, [2011] 1 S.C.R. 433, at para. 17). In the context of factual generalizations, so long as the assessment remains focused on whether there was any palpable error, such an exercise remains an integral part of the judicial function of a reviewing court.

[98]                         Once a palpable error has been identified, the reviewing court must also find that the erroneous reliance on the assumption was overriding, in that it is “shown to have affected the result” or “goes to the very core of the outcome of the case” (Clark (2005), at para. 9; Benhaim, at para. 38, citing South Yukon Forest Corp., at para. 46). If it cannot be shown that the error was palpable and overriding, a trial judge’s assessment of credibility or reliability will be entitled to deference and there will be no basis for appellate intervention.

[99]                         Given that I have found that a breach of the proposed rule against ungrounded common-sense assumptions should not be recognized as an error of law, many of the assumptions identified in the cases below, including those in the cases before this Court, in fact should have been reviewed for palpable and overriding error. Such an approach accords due deference to the trial judges’ factual findings and the role common sense played in their testimonial assessments. At the same time, it should also be emphasized that common sense is far from a catch-all phrase that licenses any form of reasoning, no matter how faulty. Common sense is not always “common”, does not always make “sense”, and worst of all, may be based on falsehoods or discriminatory beliefs. However, so long as the trial judge’s use and invocation of common sense is appropriately constrained by the legal principles applicable to appellate review in general, there is nothing inherently objectionable about its use in testimonial assessment. If and when the permissible bounds of common sense have been exceeded, appellate courts may intervene. As Fitch J.A. succinctly summarized in Pastro (at para. 41):

Judges are entitled, and expected, to rely on their life experience in making credibility findings. This necessarily includes drawing common-sense inferences from established facts. Juries are routinely instructed along the same lines — to come to common-sense conclusions based on the evidence they accept. Where it is apparent from a review of the reasons as a whole that a credibility assessment is rooted in the evidence, and is the product of a case-specific determination about what the complainant and accused did or did not do, there will be no basis for appellate intervention, absent palpable and overriding error in fact . . . . [Emphasis added; citations omitted.]

IV.         Application

[100]                     Having rejected the proposed rule, I would allow both appeals and restore the convictions. In both cases, the Court of Appeal erred in using the rule against ungrounded common-sense assumptions and reviewing the alleged improper generalizations on a correctness standard. As a result, the court failed to adopt the required contextual and functional review, leading it to overturn the trial judges’ decisions without a proper basis to do so.

[101]                     With respect, the Court of Appeal’s approach in both cases reflects the fundamental problems with the rule against ungrounded common-sense assumptions. The rule invited the court to parse the trial judges’ reasons in search of any assumptions drawn in their credibility and reliability assessments. This had the unfortunate effect of misidentifying particular phrases the trial judges used as “assumptions”, divorcing those findings from their context and the reasons as a whole. In going even further and elevating the alleged assumptions into errors of law, the court applied the wrong standard of review. As a result, in both cases, many of the assumptions identified by the court were either not assumptions at all or were otherwise ordinary factual findings reviewable for palpable and overriding error. 

A.           Mr. Kruk

[102]                     In Mr. Kruk’s case, the Court of Appeal erred in concluding that the trial judge relied on speculative reasoning in accepting the complainant’s evidence based on his observation that it “is extremely unlikely that a woman would be mistaken about that feeling [of having a penis inside her]” (trial reasons, at para. 68 (emphasis added)). In context, it is clear that the impugned statement, while perhaps unfortunately worded, was in fact not a generalization at all, but a specific articulation of the judge’s response to a theory advocated by the defence.

[103]                     First, in its analysis of the impugned statement, the Court of Appeal erred in failing to consider the whole of the judge’s findings. The court observed that the trial judge had “rejected important parts of Mr. Kruk’s evidence” — including his evidence that he had given the complainant’s parents his address, that the complainant had kicked off her own pants, and that he had only failed to contact the complainant’s mother after a certain point because his phone had died — and that the judge had found inconsistencies between Mr. Kruk’s testimony and prior statement to police (paras. 28‑29). The court also acknowledged that despite the trial judge’s concerns about the complainant’s reliability due to her intoxication, the trial judge had accepted her “core assertion” that she woke up to Mr. Kruk’s penis inside of her, as well as the circumstantial evidence in support thereof (paras. 30-32). However, these same findings appeared to play no role in the court’s analysis. Despite noting at the outset that, in addition to the complainant’s reliability, Mr. Kruk’s credibility was a central issue in the case, the court’s analysis focused solely on the issue of the complainant’s reliability, the judge’s acceptance of the complainant’s assertion of sexual assault, and the single impugned statement at para. 68 of the trial reasons: “It is extremely unlikely that a woman would be mistaken about that feeling.” The Court of Appeal characterized this single statement as the “primary reason” for the trial judge’s acceptance of the complainant’s evidence — the “why” of the judgment (paras. 63 and 65 (emphasis deleted)) — when it was para. 68 of the trial reasons as a whole, in conjunction with the judge’s adverse credibility findings concerning Mr. Kruk and his observations at paras. 69-70 of the trial reasons about circumstantial evidence consistent with a sexual encounter having occurred, that in fact conveyed “why” Mr. Kruk should be convicted.

[104]                     In this vein, the Court of Appeal also failed to give due regard to the judge’s specific credibility findings respecting Mr. Kruk. Though briefly acknowledged elsewhere, the court’s analysis of the trial judge’s errors makes no reference to the judge’s characterization of Mr. Kruk’s explanation that the complainant kicked off her own pants as “improbable in the extreme” and the finding that Mr. Kruk lied to police about an extremely important fact: that he was wearing swim trunks, with no shirt, when the complainant’s father arrived to find his daughter (trial reasons, at paras. 60 and 63). Nor is there any consideration of the fact that Mr. Kruk’s failure to contact the complainant’s mother was “inconsistent with [his stated] desire to be rid of the complainant and potentially consistent with some other motivation on his part” (trial reasons, at para. 67). These important findings are part and parcel of the judge’s finding of guilt and worked to contextualize the impugned statement about the complainant — “a woman” — not being mistaken about the feeling of penetration. In my respectful view, this statement was merely a reflection of the trial judge’s reasoning as opposed to reliance on an improper generalization.

[105]                     Viewed as a whole and in context, the trial judge did not reject the defence theory because of an assumption that no woman would be mistaken, but rather because he accepted the complainant’s testimony that she was not mistaken. Despite her intoxication, he found the complainant’s evidence on the material issue of whether there was penile-vaginal penetration to be reliable and therefore sufficient to ground a conviction. The trial judge addressed this theory at para. 68 of his reasons, acknowledging that whether the offence had been proven beyond a reasonable doubt depended on the complainant’s core assertion that she felt Mr. Kruk’s penis inside her vagina. He went on to say (at para. 68):

[The complainant’s] evidence is devoid of detail, yet she claims to be certain that she was not mistaken. She said she felt his penis inside her and she knew what she was feeling. In short, her tactile sense was engaged. It is extremely unlikely that a woman would be mistaken about that feeling.

[106]                     A functional and contextual reading of this passage of the trial reasons demonstrates that the impugned line was not an assumption or inappropriate “speculation”, as the Court of Appeal characterized it, but rather a response to the defence theory advanced in closing submissions: that the complainant, though sincere, was mistaken about the physical sensation of penile‑vaginal penetration due to her intoxication and panic after she awoke, which caused her to assume the worst.

[107]                     Even accepting the possibility that the trial judge relied on an assumption in making this finding, the Court of Appeal erred in reviewing that assumption for correctness based on the proposed rule against ungrounded common-sense assumptions. The appropriate standard of review was palpable and overriding error. In this regard, the trial judge bore in mind that, as a matter of common sense, it is extremely unlikely that someone would be mistaken about the feeling of penile‑vaginal penetration. This assumption is not wrong in the sense of being untrue or inaccurate. It was a permissible assumption against which to consider the complainant’s testimony that she was certain as to what she had felt. It discloses no palpable or overriding error.

[108]                     The Court of Appeal also considered that the trial judge’s conclusion about the complainant’s perception of penile‑vaginal penetration was not the proper subject of judicial notice, as it engaged questions of “neurology (the operation of the body’s sensory system), physiology (the impact of alcohol on perception, memory and the body’s sensory system) and psychiatry (the impact of alcohol and/or trauma on perception and memory)” (para. 67). With respect, the court’s analysis on this point illustrates the potentially absurd consequences of the rule against ungrounded common-sense assumptions. It belies belief that questions of neurology, physiology, or psychiatry would be engaged by testimony that an intoxicated witness was certain they were physically assaulted in some other way, such as a punch to the face or a kick to the shins, such that expert evidence would be required to support that testimony — yet this is exactly the type of evidence the court implied was necessary for the trial judge to reach the conclusion he did in the context of penile-vaginal penetration. Even putting aside the impracticalities associated with finding experts willing and able to testify on such issues, such testimony is simply not necessary to establish what the trial judge determined to be what happened, having heard the complainant’s testimony and considering it in light of all the other evidence.

[109]                     Where a person with a vagina testifies credibly and with certainty that they felt penile‑vaginal penetration, a trial judge must be entitled to conclude that they are unlikely to be mistaken. While the choice of the trial judge to use the words “a woman” may have been unfortunate and engendered confusion, in context, it is clear the judge was reasoning that it was extremely unlikely that the complainant would be mistaken about the feeling of penile‑vaginal penetration because people generally, even if intoxicated, are not mistaken about that sensation. In other words, the judge’s conclusion was grounded in his assessment of the complainant’s testimony. The Court of Appeal erred in finding otherwise.

B.            Mr. Tsang

[110]                     In Mr. Tsang’s case, the Court of Appeal erred in concluding that the judge’s assessment of the accused and the complainant’s credibility was fatally affected by three material unfounded assumptions about normal behaviour: a person would not ask to be spanked “out of the blue”; a controlling person would not refrain from engaging in vaginal intercourse because they could not find a condom; and a person would not abruptly drive away from someone with whom they had just had consensual sex. Again, these first two assumptions were, in fact, not assumptions but statements that reflected the trial judge’s reasoning process and findings of fact. The third assumption, though truly an assumption and one that was palpably incorrect, was not overriding, as it did not affect the core of the trial judge’s finding of guilt.

[111]                     As a general observation, just as was the case in Mr. Kruk’s appeal, the Court of Appeal erred by not considering the whole of the trial judge’s reasons when it conducted its review. At the outset of its analysis, the court stated at para. 29 that the “focus of [their] attention” was on the trial judge’s assessment of the complainant’s and Mr. Tsang’s credibility set out at paras. 117-58 of her reasons. While the identified paragraphs comprise the whole of the trial judge’s credibility analysis, the court’s assessment of each alleged error was narrower in scope. In particular, the court artificially segmented what happened in the car from what happened earlier in the night and thus did not view the judge’s analysis as a whole.

[112]                     This is especially apparent from the court’s finding that the trial judge had rejected Mr. Tsang’s evidence about what happened in the car, where the alleged assault occurred, based on unfounded assumptions. The court explicitly held that the judge’s prior finding that Mr. Tsang was untruthful about what had occurred at the after-party at the club was “bound to have coloured her assessment of [Mr. Tsang’s] evidence about what happened later”, but the court was unwilling to find that the earlier rebuffing of Mr. Tsang “weighed heavily in her rejection” of his evidence on what happened in the car (para. 45). In other words, the court disregarded the trial judge’s adverse credibility findings against Mr. Tsang (trial reasons, at paras. 118-23).

[113]                     The court also segmented the complainant’s evidence. It acknowledged the judge’s finding that the complainant had no interest in Mr. Tsang while at the club, but then characterized the “sexual foreplay” that occurred in the car as a “marked departure from her previous behaviour”, noting that the “complainant’s resistance to [Mr. Tsang’s] advances clearly diminished by the time they got to the parking lot” (para. 52). As such, the court determined that “the complainant’s conduct earlier in the evening cannot be said to provide a reliable basis for conclusions with respect to consensual behaviour in the parked car” (para. 52). Though not the main issue on appeal, I note that in doing so, the court improperly reweighed the evidence to come to its own finding about what evidence could provide a “reliable basis” for the judge’s subsequent conclusions.

[114]                     This description of the complainant’s behaviour in the car as a “marked departure” ignores the thrust of her evidence: that “she was engaging in risk management” and chose to remain in the car because she was “scantily attired” and did not feel safe walking home through a bad neighbourhood where the sex trade occurs (trial reasons, at para. 53). She decided to get into the backseat of the car to kiss Mr. Tsang, even though the tip of his penis was exposed, because: “. . . the two of them had mutual friends and she thought she had a general idea of who he was. She said a man could just grab her off the street or she could stay with a guy she thought she knew. She thought that maybe giving Mr. Tsang a bit of what he wanted by making out, meaning kissing, was a little safer” (para. 54 (emphasis added)).

[115]                     The trial judge’s analysis was alive to this aspect of the complainant’s evidence. In considering the evidence that the complainant sat on Mr. Tsang’s lap at the club, the judge quoted the complainant’s testimony that “you pick your fights” (para. 141). Then, in considering that the complainant directed Mr. Tsang to pull into a parking lot, the judge observed that the complainant “weighed her options in her drunken state and decided to stay with a man who she had not perceived to be a threat to her that evening rather than walk home dressed as she was in a neighbourhood she thought to be unsafe” (para. 143). In finding that the complainant’s behaviour was a “marked departure”, the Court of Appeal did not accord appropriate deference to the trial judge’s findings about what happened earlier in the night and the judge’s uses of those findings to assess the testimony about what happened in the car, and improperly reweighed the evidence.

[116]                     Just as in Mr. Kruk’s case, the court’s approach to reviewing the reasons led it astray. When the trial judge’s reasons are viewed holistically and in context, in my respectful view, the first two “assumptions” the Court of Appeal identified were not assumptions at all, and the third assumption did not amount to an overriding error.

[117]                     On the first alleged assumption, the Court of Appeal found two paragraphs in particular to be problematic:

While I accept Mr. Tsang’s evidence that [the complainant] got into the back seat with him and consented to engage in some sexual foreplay, it is not believable that [the complainant] would have asked him to spank her. Moreover, given the encounter thus far in the back seat, a belief that [the complainant] was gearing up for rough sex with Mr. Tsang would be a fanciful and unfounded one on his part. The comments he testified to sounded contrived as if he were describing how he thought such an encounter might unfold rather than what really happened in the parking lot that night.

 

. . .

 

. . . Mr. Tsang’s evidence this was consensual rings hollow in the opinion of the Court. His testimony about the things he said to [the complainant] and what she said to him seemed lifted from a pornographic script completely at odds with their encounter up to that point.

 

(trial reasons, at paras. 126 and 148)

[118]                     The Court of Appeal characterized the trial judge’s conclusion that it was unbelievable the complainant would ask to be spanked as “a bald statement untethered to any evidence”, considering that the trial judge’s “primary rationale” for rejecting the conclusion appeared to be “simply that it is ‘unbelievable’” (paras. 48-49). The court went on to emphasize that, absent extrinsic evidence to support it, the trial judge’s conclusion about spanking could “only have been founded upon an assumption about what activity [the complainant] might have willingly engaged in after she willingly engaged in some sexual foreplay”, and that the trial judge’s assessment of the evidence in this regard was thus affected by “implicit, unsupported assumptions about ‘normal behaviour’” (para. 53).

[119]                     While the Court of Appeal appeared to be preoccupied with finding some evidence in the record to support the trial judge’s conclusion, it simultaneously disregarded the fact that that the trial judge’s assessment of the complainant’s evidence was based on a record that, when assessed as a whole, contained the very evidence needed to contextualize that conclusion. Although clearly there was no evidence about what activities the complainant would have willingly engaged in after sexual foreplay, the complainant specifically testified that she had not agreed to being spanked and had not asked the accused for rough sex. This evidence formed the backdrop for the credibility findings at paras. 117-58 of the trial judge’s reasons and for her ultimate decision to reject the accused’s account as contrived. As was the case in Mr. Kruk’s appeal, the Court of Appeal’s reasoning in this context reflects the fundamental problems with the proposed rule against ungrounded common-sense assumptions: in search of some extrinsic source of support for generalizations, appellate courts risk missing the evidence that is actually in the record. 

[120]                     The trial judge’s conclusion that it was unbelievable the complainant would ask to be spanked or for rough sex must also be understood in the context of the broader body of evidence that the complainant was uninterested in Mr. Tsang, but engaged in a risk analysis and agreed to limited sexual activity in the car — kissing, but nothing more. The trial judge’s finding in this respect was tethered to her assessment of Mr. Tsang’s evidence as contrived: her remark that Mr. Tsang sounded “as if he were describing how he thought such an encounter might unfold” represents her explanation for why she found Mr. Tsang’s testimony that the complainant asked to be spanked to be unbelievable (para. 126). Put simply, the judge accepted the complainant’s evidence that she did not ask to be spanked or consent to have sex — rough or otherwise — and rejected Mr. Tsang’s evidence. Her conclusion was not based on an assumption, but rather represents her assessment of the direct testimony of the two key witnesses.

[121]                     With respect to the second alleged assumption, the Court of Appeal held the trial judge erred in finding that Mr. Tsang’s description “about the prospect of intercourse being thwarted by the lack of a condom in his car when there was one available was contrived . . . and contrary to the level of control he conveyed about that evening and in court” (trial reasons, at para. 129). The Court of Appeal considered this finding to be “problematic” and characterized it as an assumption that a controlling person would not refrain from having unprotected sex (paras. 64-65).

[122]                     Mr. Tsang’s evidence was not assessed in relation to an assumption about whether controlling people would refrain from sex without a condom — rather, his evidence on this point was inconsistent with the judge’s other findings, and so was disbelieved. Properly viewed, the judge’s finding was that the accused himself was controlling and did not respect the complainant’s wishes, a finding which the Court of Appeal agreed was open to her to make. In particular, the judge found that Mr. Tsang had, at various points, exercised control to ignore the wishes of the complainant and her friend, including by insisting that the women drink at the club and leave their bags in his car. In light of these findings, the judge found Mr. Tsang’s evidence about not having sex, because the complainant asked if he had a condom and he could not find one, to be unbelievable. Because the Court of Appeal artificially isolated the findings about what happened in the car from what happened in the club, it did not properly consider these other findings and thus reached the erroneous conclusion that the judge had made an assumption.

[123]                     The third assumption the Court of Appeal identified relates to the judge’s assessment of Mr. Tsang driving away quickly after dropping the complainant off. The judge made the following remarks:

His lack of interest . . . at [the complainant’s] invitation to meet again was at odds with his evidence that they had just had a great time, but is consistent with a non-consensual event where he got what he wanted without regard for her and drove away.

 

. . .

 

Mr. Tsang’s testimony aligned with that of [the complainant] and [the complainant’s friend] that he drove off as soon as [the complainant] got out of his car and he did not watch her go inside the house. I find this fact more consistent with [the complainant’s] claim of non-consensual sex than with Mr. Tsang’s version of what had just happened. He took off right away because of what he had just done to her and because she meant nothing to him. It is inconsistent with his evidence that he wanted to pleasure [the complainant] that night. [paras. 131 and 153]

[124]                     I agree that the judge erred here in relying on an assumption that people do not drive off quickly after consensual sexual encounters. However, without the rule against ungrounded common-sense assumptions in play, this assumption again falls to be assessed under the palpable and overriding error standard. This is the standard the Court of Appeal should have applied, and had it done so, it would have found no basis for reviewable error.

[125]                     I accept that the inaccuracy of this assumption is palpable, as it is obviously untrue and plainly observable. Mr. Tsang’s after-the-fact decision to drive off quickly is neither consistent nor inconsistent with a sexual assault having occurred. As a matter of logic, the speed at which someone leaves after sexual activity generally has no bearing whatsoever on whether the activity was consensual or non-consensual. One can think of plenty of reasons why a person such as Mr. Tsang would want or need to depart quickly after consensual sex — including reasons completely irrelevant to the sexual activity itself, such as needing to get home after a late night. There are also plenty of reasons why a person who commits sexual assault would insist on staying at the scene for longer once the assault was over — for example, to ensure that the complainant does not say anything to others about what happened. In the context of this case, the fact that Mr. Tsang drove off as soon as the complainant got out of his car and “did not watch her go inside the house” (para. 153) has nothing to do with whether the complainant consented to the sexual activity earlier that night. I accept that the trial judge’s reliance on this generalization was a palpable error.

[126]                     Nevertheless, the assumption is not overriding. Although it formed part of the judge’s assessment of Mr. Tsang’s testimony and contributed to the trial judge’s finding that Mr. Tsang was a non-credible witness who “gave a very contrived and unbelievable story” (para. 156), the assumption itself did not form an essential part of the trial judge’s decision to convict Mr. Tsang. Given all the other adverse credibility findings against Mr. Tsang, I am left with no doubt whatsoever that the judge would still have disbelieved him even without the assumption in play. Moreover, given the judge found the complainant to be credible and reliable, I am left with no doubt that the judge would still have convicted Mr. Tsang of sexual assault.

V.           Disposition

[127]                     I would allow both appeals, set aside the orders of the British Columbia Court of Appeal, and restore the convictions.

                   The following are the reasons delivered by

                   Rowe J. —

TABLE OF CONTENTS
Paragraph
I.      Overview	128
II.    Jurisprudential Background	134
III.   Principles of the Fact-Finding Process	141
A.    Foundational Concepts	142
B.    Generalized Expectations in the Fact-Finding Process	151
IV.  Framework for Appellate Review of the Use of Generalized Expectations in the Fact-Finding Process	160
A.    Did the Trial Judge Rely on a Generalized Expectation in Their Reasoning Process?	165
(1)      Materiality and the Curative Proviso	171
B.    If the Trial Judge Relied on a Generalized Expectation, Was the Expectation Reasonable?	181
(1)      Question of Law or Fact?	190
(2)      The “Rules” in J.C.	209
C.    Did the Trial Judge Rely on a Generalized Expectation as Itself a Conclusive and Indisputable Fact?	214
V.    Application to These Appeals	220
A.    Mr. Kruk	221
B.    Mr. Tsang	233
(1)      A Person Would Not Ask To Be Spanked While Engaging in Sexual Foreplay “Out of the Blue”	236
(2)      A Controlling Person Would Not Refrain From Engaging in Vaginal Intercourse Because of the Absence of a Condom	240
(3)      A Person Would Not Abruptly and Unceremoniously Leave Another With Whom He Had Engaged in Consensual Sex	245
VI.   Conclusion	249
Appendix

I.               Overview

[128]                     These two appeals ask how appellate courts should review trial judges’ reliance on generalized expectations based on common sense and human experience in the fact-finding process. In both cases, the accused was convicted at trial by judge alone of one count of sexual assault, but the Court of Appeal for British Columbia unanimously ordered a new trial on the basis that the trial judge erred in their reliance on generalized expectations about human perception or behaviour. The Crown appeals to this Court, seeking to restore the convictions.

[129]                     Generalized expectations based on common sense and human experience play a necessary role in the judicial fact-finding process. They serve as a logical benchmark against which to compare the evidence for the purposes of drawing inferences from circumstantial evidence or assessing a witness’s credibility. However, intermediate appellate courts have increasingly identified concerns about the limits of this exercise. In these reasons, I propose three questions that an appellate court should ask when reviewing for potential legal error in a judge’s reliance on generalized expectations in the fact-finding process.

[130]                     First, did the trial judge rely on a generalized expectation in their reasoning process? Often, a factual conclusion may appear to reflect a generalized expectation, but the judge may not have actually relied on such an expectation and instead assessed the evidence with reference to other accepted evidence or facts from the trial. If the appellate court determines that the judge relied on a generalized expectation, then no error of law has yet been identified, and the analysis will proceed to the second question. In contrast, where a judge has not relied on any generalized expectation, the review for potential error under this framework ends.

[131]                     Second, if the trial judge relied on a generalized expectation, was the expectation reasonable? In my view, it is an error of law for a judge to rely on an unreasonable generalized expectation to assess the evidence. Unreasonable generalized expectations masquerading as common sense or collective human experience are not a legitimate basis on which to assess and understand the evidence in a criminal trial. They are not confined to what might be considered “myths” or “stereotypes” or to the context of sexual assault trials, but may arise throughout the fact-finding process in any criminal trial. Public confidence in the administration of justice and the judicial fact-finding process require that appellate courts be able to intervene where trial judges employ generalized expectations that are not a reasonably accurate reflection of what is true in most circumstances and are not a reliable benchmark to assess the evidence.

[132]                     Third, did the judge rely on a generalized expectation as itself a conclusive and indisputable fact? Although judges have considerable latitude to rely on reasonable generalized expectations as a benchmark for assessing the evidence, such expectations are not a replacement for the evidence. It is an error of law for a judge to fail to consider all of the evidence on the ultimate issue of guilt or innocence, or to make a factual conclusion in the absence of evidence. Where a judge relies on a generalized expectation as itself a factual conclusion, what the judge is really doing is taking judicial notice, which is subject to a stringent test.

[133]                     Applying these principles to the appeals at hand, I would allow the appeals and restore the convictions. In Mr. Kruk’s case, the trial judge relied on a generalized expectation about the likelihood of a woman being mistaken about an invasive physical experience (2020 BCSC 1480). This was a reasonable expectation about general human perception. The trial judge did not treat this as an indisputable fact but instead used it as a benchmark to assess the complainant’s evidence in light of the totality of the evidence. There was therefore no basis for the Court of Appeal to intervene. In Mr. Tsang’s case, the Court of Appeal wrongly identified the trial judge as having relied on a generalized expectation in two instances where the latter was actually assessing the whole of the evidence; this led the Court of Appeal to move beyond its role and reweigh the evidence. In one instance, the trial judge did err by relying on an unreasonable expectation about how people ordinarily behave after a consensual sexual encounter. However, this error could have had no impact on the verdict, and I would have maintained the conviction pursuant to the curative proviso under s. 686(1)(b)(iii) of the Criminal Code , R.S.C. 1985, c. C-46 .

II.            Jurisprudential Background

[134]                     The two decisions on appeal are part of an influx of recent decisions by intermediate appellate courts on the use of “common sense”, “human experience”, “logic”, “generalizations”, “assumptions”, “myths”, “stereotypes”, and other similar concepts used by trial judges in the fact-finding process. Such issues have become especially prevalent in appeals from sexual assault verdicts, both from acquittals and convictions (see, e.g., R. v. Kodwat, 2017 YKCA 11; R. v. A.R.D., 2017 ABCA 237, 422 D.L.R. (4th) 471 (“A.R.D.”), aff’d R. v. A.R.J.D., 2018 SCC 6, [2018] 1 S.C.R. 218 (“A.R.J.D.”); R. v. Quartey, 2018 ABCA 12, 430 D.L.R. (4th) 381, aff’d 2018 SCC 59, [2018] 3 S.C.R. 687; R. v. Kiss, 2018 ONCA 184; R. v. L. (J.), 2018 ONCA 756, 143 O.R. (3d) 170; R. v. Paulos, 2018 ABCA 433, 79 Alta. L.R. (6th) 33; R. v. A. (A.B.), 2019 ONCA 124, 145 O.R. (3d) 634; R. v. F.B.P., 2019 ONCA 157; R. v. Cepic, 2019 ONCA 541, 376 C.C.C. (3d) 286; R. v. Pilkington, 2019 BCCA 374; R. v. Percy, 2020 NSCA 11, 61 C.R. (7th) 7; R. v. Roth, 2020 BCCA 240, 66 C.R. (7th) 107; R. v. Delmas, 2020 ABCA 152, 452 D.L.R. (4th) 375, aff’d 2020 SCC 39, [2020] 3 S.C.R. 780; R. v. Pastro, 2021 BCCA 149, 71 C.R. (7th) 296). Various attempts have been made to articulate the principles by which appellate courts should approach these issues. A coherent solution has been elusive.

[135]                     The parties and interveners point to the efforts of the Court of Appeal for Ontario in R. v. J.C., 2021 ONCA 131, 401 C.C.C. (3d) 433, with varying degrees of approval. This decision was noted to be helpful by the Court of Appeal for British Columbia in Mr. Tsang’s appeal (2022 BCCA 345, 419 C.C.C. (3d) 187). It was not cited directly in the decision on appeal in Mr. Kruk’s case (2022 BCCA 18), but both of these decisions refer to similar authorities, such as Roth, Cepic, and R. v. Perkins, 2007 ONCA 585, 223 C.C.C. (3d) 289. It is therefore helpful to review J.C. at some length.

[136]                     In J.C., Justice Paciocco, for the Court of Appeal, described “two relevant legal rules that identify impermissible reasoning relating to the plausibility of human behaviour” (para. 57). First, there is a “rule against ungrounded common-sense assumptions”, which says that “judges must avoid speculative reasoning that invokes ‘common-sense’ assumptions that are not grounded in the evidence or appropriately supported by judicial notice” (para. 58). The rule “does not bar using human experience about human behaviour to interpret evidence”; however, “[i]t prohibits judges from using ‘common-sense’ or human experience to introduce new considerations, not arising from evidence, into the decision-making process, including considerations about human behaviour” (para. 61).

[137]                     Second, there is an overlapping “rule against stereotypical inferences”, which says that “factual findings, including determinations of credibility, cannot be based on stereotypical inferences about human behaviour” (para. 63). This means “it is an error of law to rely on stereotypes or erroneous common-sense assumptions” about how a complainant or accused person is expected to act, either to bolster or compromise their credibility (ibid.). This rule only prohibits inferences based on “stereotypes” or “prejudicial generalizations” from being drawn, not the admission or use of certain kinds of evidence. A factual conclusion may logically reflect a stereotype but will not be an error if it is based on the evidence.

[138]                     In addition, Justice Paciocco considered that “errors” arising from violating the rule against ungrounded common-sense assumptions or the rule against stereotypical inferences “are reversible only when they ‘ground’ the relevant inference by playing a material or important role in the impugned conclusion” (para. 71).

[139]                     J.C. has been relied on by intermediate appellate courts in numerous instances in the few years since it was decided. This reflects the prevalence of the issues now before this Court concerning trial judges’ reliance on matters such as “common sense” and “human experience” to decide cases. Professor Lisa Dufraimont describes the attempt to clarify the law in J.C. as a “real advance”, while expressing some hesitation with respect to the breadth and malleability of the described rules (“Current Complications in the Law on Myths and Stereotypes” (2021), 99 Can. Bar Rev. 536, at pp. 563-64).

[140]                     A key issue raised by some of the parties and interveners in these appeals is whether the “rule against ungrounded common-sense assumptions” described in J.C. should be recognized as a basis for appellate intervention. The Crown and certain interveners say that the rule is unworkable and should be rejected by this Court, while the respondents and other interveners support the proposed approach in J.C. In contrast, the “rule against stereotypical reasoning” is (seemingly) less controversial; the parties all agree that stereotypical reasoning is a basis for appellate intervention. There is, however, a lack of clarity as to how some of the parties and interveners view the relationship between these two “rules”. Further, the parties disagree as to whether an error must play a “material” or “important” role in the impugned conclusion in order for the error to be “reversible”, as described in J.C.

III.         Principles of the Fact-Finding Process

[141]                     The proliferation of appellate jurisprudence and the submissions before this Court point to the need for guidance in the form of a clear and consistent framework for appellate review. In my view, a significant reason for the lack of clarity and consistency is the imprecise use of terminology, such as “common sense”, “stereotypes”, “myths”, “inferences”, “judicial notice”, and “speculation”. I therefore find it helpful to begin with what I consider to be fundamental principles of the fact-finding process. I set out these principles with a view to using precise terms deliberately.

A.           Foundational Concepts

[142]                     The essential purpose and feature of the trial process is the search for truth by identifying the “fact[s]” to which the law will be applied (S. N. Lederman, M. K. Fuerst and H. C. Stewart, Sopinka, Lederman & Bryant: The Law of Evidence in Canada (6th ed. 2022), at ¶1.40). In a criminal trial, the trier of fact is tasked with applying the law to the facts to determine whether the accused is guilty or not guilty of the offences charged. The law is set out by statute or by common law and must be correctly understood and applied. The facts, on the other hand, are the province of the trier of fact. The trier of fact’s primary task is to make conclusions on the “facts in issue” in the trial; in a criminal case, the required elements of the offence are the key facts in issue. In addition to the facts in issue, any other fact will be relevant where it “proves or renders probable the past, present or future existence (or non-existence) of any fact in issue” (¶2.57). In addition, “[f]acts relating to the credibility of the witness giving direct or circumstantial evidence of a fact in issue are relevant” (¶2.58).

[143]                     A conclusion of fact can be arrived at in one of two general ways: (a) by taking of judicial notice or (b) by “findings of fact” derived from the evidence or through “inferences” from other facts.

[144]                     “Judicial notice” is the acceptance by a court of the truth of a particular fact without “proof” (Lederman, Fuerst and Stewart, at ¶19.25). The threshold for taking judicial notice of fact is strict — “a court may properly take judicial notice of facts that are either: (1) so notorious or generally accepted as not to be the subject of debate among reasonable persons; or (2) capable of immediate and accurate demonstration by resort to readily accessible sources of indisputable accuracy” (R. v. Find, 2001 SCC 32, [2001] 1 S.C.R. 863, at para. 48). Once a fact is taken on judicial notice, consideration of evidence on the fact will generally be foreclosed, as satisfying the test for judicial notice means the fact is not reasonably debatable; however, this does not mandate that any “inference” from the judicially-noticed fact must be drawn (Lederman, Fuerst and Stewart, ¶¶19.74-19.75). Due to the consequences of drawing a fact on judicial notice, judges should give the parties notice and the opportunity to respond before relying on such facts (¶¶19.78 et seq.).

[145]                     “Evidence” is the primary means to facilitate the introduction of all logically relevant facts at trial (Lederman, Fuerst and Stewart, at ¶1.1). There are various types of evidence; I focus in these reasons on viva voce testimony from witnesses at trial. Evidence must be admissible in order to be considered by the trier of fact; I do not discuss issues of admissibility in these reasons. Evidence is not itself fact. Witnesses testify as to their observations and experiences, and the trier of fact may accept all, some, or none of the witness’s evidence as fact (see, e.g., R. v. François, [1994] 2 S.C.R. 827, at p. 837; R. v. D.A.I., 2012 SCC 5, [2012] 1 S.C.R. 149, at para. 72).

[146]                     Whether evidence is accepted as fact by the trier of fact will depend on an assessment of the witness’s “credibility” and “reliability”. Credibility refers to a witness’s honesty or sincerity. Reliability, meanwhile, is about the accuracy of the witness’s testimony, referring to the witness’s ability to observe, recall, and recount events (see R. v. H.C., 2009 ONCA 56, 244 O.A.C. 288, at para. 41). Assessing credibility and reliability is not a science (see R. v. Gagnon, 2006 SCC 17, [2006] 1 S.C.R. 621, at para. 20). Credibility and reliability are assessed based on various factors including the character, demeanor, and conditions and capabilities of the witness, the plausibility and internal consistency of the testimony, and supporting information; it is also assessed in light of its consistency with other facts and evidence (D. M. Paciocco, P. Paciocco and L. Stuesser, The Law of Evidence (8th ed. 2020), at p. 593).

[147]                     There are, broadly speaking, two types of evidence: “direct evidence” and “circumstantial evidence”. Direct evidence is “evidence which, if believed, resolves a matter in issue” (R. v. Cinous, 2002 SCC 29, [2002] 2 S.C.R. 3, at para. 88, quoting D. Watt, Watt’s Manual of Criminal Evidence (2001), at § 8.0). For example, a witness who testifies that they saw it raining outside is providing direct evidence that it was raining. Circumstantial evidence, also known as “[i]ndirect evidence”, is evidence from which the trier of fact is asked to draw certain “inferences” (R. v. Villaroman, 2016 SCC 33, [2016] 1 S.C.R. 1000, at para. 23, quoting National Committee on Jury Instructions of the Canadian Judicial Council, Model Jury Instructions (online), s. 10.2). For example, a witness who testifies they saw someone enter the courthouse wearing a raincoat and carrying an umbrella, both dripping wet, is providing circumstantial evidence, from which an inference may be drawn that it was raining (see ibid.). Circumstantial evidence “is all about inferences” (S. C. Hill, D. M. Tanovich and L. P. Strezos, McWilliams’ Canadian Criminal Evidence (5th ed. (loose-leaf)), at § 31:17, citing R. v. Gibson, 2021 ONCA 530, 157 O.R. (3d) 597, at para. 76).

[148]                     An “inference” is a finding of fact that may logically and reasonably be drawn from another fact (or group of facts) found or otherwise established in the proceedings (e.g., through judicial notice) (Hill, Tanovich and Strezos, at § 31:17, citing R. v. Chanmany, 2016 ONCA 576, 338 C.C.C. (3d) 578, at para. 45; Lampard v. The Queen, [1969] S.C.R. 373; D. Watt, Watt’s Manual of Criminal Evidence (2023), at §12.01). Thus, triers of fact draw inferences from facts (known as “primary facts”) in order to conclude the existence of further facts (which are inferences). Inferences may — not must — be drawn in the circumstances (Watt, at §12.01).

[149]                     An “opinion” can be understood as a particular inference proffered by a witness (see White Burgess Langille Inman v. Abbott and Haliburton Co., 2015 SCC 23, [2015] 2 S.C.R. 182, at para. 14). The line between fact and opinion is not always clearly drawn (Graat v. The Queen, [1982] 2 S.C.R. 819, at p. 835). As a general rule, however, witnesses may not give opinions, but should testify only to “facts” in their knowledge, observation, and experience; it is for the trier of fact to draw inferences from proven facts (Lederman, Fuerst and Stewart, at ¶12.2). However, a qualified expert witness may provide the trier of fact with a “ready-made inference” that the trier of fact would otherwise not be able to draw because of the technical nature of the subject matter (ibid.; R. v. Mohan, [1994] 2 S.C.R. 9). A lay witness may also be able to give an opinion on certain matters (see Graat); however, as the evidence approaches the central issues that the court must decide, resistance to admissibility increases (Lederman, Fuerst and Stewart, at ¶12.15).

[150]                     “Speculation” is a concept that has been given several meanings in different contexts. Notably, inferences must be ones which can be reasonably and logically drawn from a primary fact established by the evidence (or on judicial notice); however, “[a]n inference which does not flow logically and reasonably from established facts cannot be made and is condemned as conjecture and speculation” (R. v. Morrissey (1995), 22 O.R. (3d) 514 (C.A.), at p. 530). It follows that reasoning that is not based on the facts established by evidence is generally speculative. However, “[t]he boundary that separates permissible inference from impermissible speculation in connection with circumstantial evidence is often a very difficult one to determine” (Watt, at §12.01; see also Hill, Tanovich and Strezos, at § 31:17; Canadian Pacific Railway Co. v. Murray, [1932] S.C.R. 112, at p. 117).

B.            Generalized Expectations in the Fact-Finding Process

[151]                     The foregoing sets out certain foundational concepts of the fact-finding process. Missing from this array is the role of the trial judge’s reasoning, which bridges the gaps between evidence and fact, primary fact and inference, and so on.

[152]                     Common sense has been described as “both self-evident and inscrutable” (P. Cochran, Common Sense and Legal Judgment: Community Knowledge, Political Power, and Rhetorical Practice (2017), at p. 15). At various points in the fact-finding process, trial judges will necessarily rely on their “common sense” and “human experience” (or “experience of human affairs” (R. v. Béland, [1987] 2 S.C.R. 398, at p. 418)) when assessing the evidence and deciding the facts in a case (see R. v. S. (R.D.), [1997] 3 S.C.R. 484, at paras. 13, 29, 38 and 130; Villaroman, at para. 30). As Justice Major explained in S. (R.D.):

     It is apparent, and a reasonable person would expect, that triers of fact will be properly influenced in their deliberations by their individual perspectives on the world in which the events in dispute in the courtroom took place. Indeed, judges must rely on their background knowledge in fulfilling their adjudicative function. As David M. Paciocco and Lee Stuesser write in their book The Law of Evidence (1996), at p. 277: 

 

In general, the trier of fact is entitled simply to apply common sense and human experience in determining whether evidence is credible and in deciding what use, if any, to make of it in coming to its finding of fact. [Emphasis deleted; para. 39.]

[153]                     Thus, common sense and human experience are the foundation for what can be described as “generalized expectations” — whether about typical human behaviour, human perception, or other shared experiences of humanity. In these reasons, I will rely on the language of “generalized expectations” to encompass considerations based on “common sense” or “human experience” in the judicial reasoning process. This phrase highlights that such considerations are both generalized (meaning they relate not to any particular person or event, but about people or things in general) and are expectations (meaning they are ideas about what is usually or ordinarily true, not about what is definitively true in a particular case). Such generalized expectations are often referred to, inter alia, as “assumptions” (including “common-sense assumptions”), “generalizations”, or “common-sense inferences”. I do not favour those terms, insofar as they necessarily imply error through negative connotations or invite confusion with inferences as factual conclusions drawn from the evidence or other primary facts.

[154]                     Common sense, upon which generalized expectations are based, is the “predominant source of assessment for trier of fact decision-making” (Hill, Tanovich and Strezos, at § 31:16). Generalized expectations are an input into the reasoning process, acting as a logical benchmark against which to compare the evidence. For example, during the reasoning process used to arrive at an inference from a primary fact, trial judges may compare a proposed inference against the benchmark of a generalized expectation of events based on common sense and human experience. To parse out the wet raincoat and umbrella example raised earlier, the proposed inference might be that it was raining. A witness may offer the circumstantial evidence that they saw someone with a dripping raincoat and umbrella. In order to draw the inference that it was raining, the trier of fact might rely on the generalized expectation that raincoats and umbrellas are usually wet due to the rain. My point is that the generalized expectation is an input in the inference-drawing process, whether implicit or explicitly relied on. The inference, of course, is subject to the other evidence in the record.

[155]                     Similarly, in assessing the credibility of a witness, trial judges are also expected to apply common sense and human experience as a benchmark against which to weigh the plausibility of the evidence (Paciocco, Paciocco and Stuesser, at p. 608; see also R. v. Marquard, [1993] 4 S.C.R. 223, at p. 248); thus, the trial judge applies their experiences and common sense when judging the witness’s demeanour, character, or the internal or external plausibility of their evidence to decide whether an individual’s evidence is believable (Paciocco, Paciocco and Stuesser, at p. 593).

[156]                     Generalized expectations based on common sense are frequently recognized in the jurisprudence. For example, it is a common-sense expectation that a person foresees the natural and probable consequences of their actions (R. v. Walle, 2012 SCC 41, [2012] 2 S.C.R. 438, at para. 64), or is unlikely to flee a crime scene if they are not responsible for the act (R. v. Jacquard, [1997] 1 S.C.R. 314, at para. 50), or is unlikely to make false statements against their own interest (R. v. O’Brien, [1978] 1 S.C.R. 591, at p. 594). There is no closed list of common-sense assumptions; these are merely examples of the basic premise that triers of fact need not check their common sense and life experiences at the courthouse door.

[157]                     Some generalized expectations have been recognized in the jurisprudence as myths and stereotypes. “[M]yths and stereotypes” as a unified phrase is often understood in Canadian legal discourse as referring to “false beliefs about sexual assault that distort the fact-finding process”, often concerning complainants and particularly women and children (L. Dufraimont, “Myth, Inference and Evidence in Sexual Assault Trials” (2019), 44 Queen’s L.J. 316, at pp. 330-32). This Court has recognized “the prevailing existence of such myths and stereotypes” based on overwhelming evidence from relevant social science literature (Find, at para. 101). Regarding complainants in cases of sexual assault, this Court has noted that myths and stereotypes “are particularly invidious because they comprise part of the fabric of social ‘common sense’ in which we are daily immersed” and they “create the risk that victims of abuse will be blamed or unjustly discredited in the minds of both judges and jurors” (para. 103).

[158]                     In my view, concepts such as “myths” and “stereotypes” should be given distinct meanings. I would adopt a conception of “stereotypes” as referring to biases arising from “traits that one associates with a particular group” (R. v. Chouhan, 2021 SCC 26, [2021] 2 S.C.R. 136, at para. 53 (emphasis added), per Moldaver and Brown JJ., quoting A. Roberts, “(Re)forming the Jury: Detection and Disinfection of Implicit Juror Bias” (2012), 44 Conn. L. Rev. 827, at p. 833). In contrast, a “myth” might be considered a “widely held misconception” (Oxford English Dictionary (online)) about people in general, particularly with regard to sexual behaviour; for example, the “twin myths” set out in s. 276(1) relate to all complainants where their past sexual activity is at issue, although these myths have historically and continue to be targeted disproportionately at women and children. While there may be overlap in the meaning of these terms, efforts should be made to clearly set out the relevant concerns at play in a given case.

[159]                     I note that reliance on generalized expectations in judicial reasoning has been understood by some as a form of implicit “judicial notice” (see, e.g., Hill, Tanovich and Strezos, at § 26:2). In my view, it is best to avoid the term “judicial notice” in this context to distinguish (a) the reasoning process relating to evidence based on consistency with a generalized expectation from (b) the conclusion of indisputable or notorious facts through judicial notice (as an alternative to findings of fact from the evidence). As I will explain later, these concepts should be kept distinct.

IV.         Framework for Appellate Review of the Use of Generalized Expectations in the Fact-Finding Process

[160]                     The principles described above disclose two overarching ideas, which can be in tension, though not necessarily in conflict.

[161]                      On one hand, triers of fact necessarily rely on generalized expectations based on common sense or human experience as a logical benchmark when assessing the evidence. This occurs in at least two ways: (1) when drawing inferences from primary facts (by comparing the proposed inference to a generalized expectation of what is likely to happen from a given set of circumstances), and (2) when assessing credibility (by comparing the witness’s evidence to a generalized expectation of what people do or feel).

[162]                      On the other hand, reliance on generalized expectations in a criminal proceeding is not without limits. Some expectations may not be accurate or reliable predictors of general human behaviour. This includes (but, as I will explain, is not limited to) “stereotypes” or other prejudicial assumptions about a particular group of people, as well as certain widely-held misconceptions about human behaviour that the law has identified as “myths”. Furthermore, even if an expectation may be acceptable as a general proposition, it is always possible that reality may run contrary to that expectation. The trial judge’s duty is to decide what really happened, based on the evidence. Relatedly, a factual conclusion derived entirely from a generalized expectation in the absence — or to the exclusion — of the evidence is best conceptualized as a fact drawn on judicial notice, which is subject to a stringent test. This is distinguishable from the use of generalized expectations as an input into the reasoning process when assessing evidence.

[163]                     In my view, these overarching ideas disclose three questions that an appellate court should ask when reviewing for potential legal error in a trial judge’s reliance on generalized expectations in the fact-finding process:

A. Did the trial judge rely on a generalized expectation in their reasoning process?

 

B. If the trial judge relied on a generalized expectation, was the expectation reasonable?

 

C. Did the trial judge rely on a generalized expectation as itself a conclusive and indisputable fact?

[164]                     The first question is a means of distinguishing the judge’s overall appreciation of the evidence from the use of generalized expectations in the reasoning process. The second and third questions can disclose errors of law by the trial judge, but only if, under the first question, the judge has indeed relied on a generalized expectation. I will discuss each question in turn. I include a diagram of this framework as an appendix to these reasons.

A.           Did the Trial Judge Rely on a Generalized Expectation in Their Reasoning Process?

[165]                     Where a trial judge makes a conclusion about a witness’s credibility or an inference or a finding of fact from the evidence, that conclusion may appear to reflect a generalized expectation about the behaviour of most people (or a particular type of person), but the judge may not have actually relied on a generalized expectation. Instead, the judge may have assessed the evidence by comparing it to other accepted evidence or facts from the trial. Generally speaking, an appellate court may not interfere with the trial judge’s overall appreciation of the evidence, absent a palpable and overriding error of fact or unless the resulting verdict is unreasonable under s. 686(1)(a)(i).

[166]                     Thus, where a judge is alleged to have improperly relied on a generalized expectation, appellate courts must first review the judge’s reasons to determine whether the judge indeed relied on the alleged generalized expectation. Where no generalized expectation is employed, there is no basis for further appellate scrutiny, absent some other recognized basis for appellate intervention such as an unreasonable verdict. If, on the other hand, the appellate court concludes that the trial judge did rely on a generalized expectation, then the analysis proceeds to the next question.

[167]                     At this stage of the analysis, the appellate court is not identifying any error. It is not an error of law per se for a judge to rely on a generalized expectation as a logical benchmark when assessing the evidence. To the contrary, as I have explained, this is a well-recognized and necessary part of the judicial fact-finding process. Rather, in answering this first question, the appellate court is simply determining what the trial judge really decided, why the judge decided that way, and whether there is a basis for further scrutiny.

[168]                     Many errors often alleged before appellate courts would be resolved under this question. For example, in Quartey, the accused — a man — testified that he was not interested in sex with the complainant and that he had refused the complainant’s attempt to perform fellatio on him because he did not enjoy fellatio. The trial judge rejected his evidence and found him guilty of sexual assault. On appeal, the accused argued that the judge found his testimony unbelievable because the judge assumed men are more interested in engaging in sex than women and that all men enjoy fellatio. This Court agreed with the Court of Appeal of Alberta that the trial judge did not “err by applying generalizations and stereotypes in rejecting the appellant’s evidence”, because “the trial judge’s statements in this regard were directed to the appellant’s own evidence and to the believability of the appellant’s claims about how he responded to the specific circumstances of this case, and not to some stereotypical understanding of how men in those circumstances would conduct themselves” (para. 3 (emphasis in original); see also, e.g., Pastro, at paras. 41 and 66; Percy, at para. 107).

[169]                     As Professor Dufraimont observes, “[t]he clear implication of Quartey . . . is that the accused’s appeal would have succeeded if the trial judge had indeed relied on stereotypes about men and male sexuality” ((2021), at p. 548). Thus, a negative answer to this first question may obviate appellate intervention, even where an alleged generalized expectation, had it been relied on, would have amounted to an error of law (as discussed below).

[170]                     How can appellate courts determine whether a trial judge relied on a generalized expectation (meaning the analysis proceeds to the next two questions below) or made a determination based on the evidence and facts in the particular case (meaning there is no need for further scrutiny under this framework)? This is a highly case-dependent inquiry. However, this Court has provided repeated guidance in recent cases to which appellate courts should direct themselves. An appellate court must assess whether the reasons, read as a whole and in the context of the live issues at trial, explain what the trial judge decided and why they decided that way in a manner that permits effective appellate review. Appellate courts must not finely parse the trial judge’s reasons in a search for error (see R. v. G.F., 2021 SCC 20, [2021] 1 S.C.R. 801, at para. 69, referring to R. v. Chung, 2020 SCC 8, [2020] 1 S.C.R. 405, at paras. 13 and 33). Thus, appellate courts must not view a single statement that might appear to reflect a generalized expectation in isolation. The reasons read as a whole and considered in light of the evidence and submissions at trial may indicate that the statement related to the circumstances of the particular case or a conclusion made by the judge elsewhere in the reasons.

(1)          Materiality and the Curative Proviso

[171]                     Before turning to the second question of my proposed framework, I pause to distinguish the question of whether the trial judge relied on a generalized expectation, as described above, from the question of “materiality” and the operation of the curative proviso under s. 686(1)(b)(iii) of the Criminal Code . As I have explained, the Court of Appeal for Ontario in J.C. considered that, after identifying errors in the trial judge’s reasons (with respect to a “rule against ungrounded common-sense assumptions” or a “rule against stereotypical reasoning”), an appellate court may reverse the trial judge’s decision “only when [the errors] ‘ground’ the relevant inference by playing a material or important role in the impugned conclusion” (para. 71). Justice Paciocco stated that the burden to prove the materiality of an error falls on the accused who appeals, and that this consideration is distinct from the operation of the curative proviso, where the burden falls on the Crown (paras. 100-101). This aspect of J.C. led the Court of Appeal for British Columbia in Mr. Tsang’s case to conduct an extensive review of the evidence and the trial judge’s reasons to determine whether the judge committed a reversible error, after having found that the judge had erred.

[172]                     The requirement for materiality or importance in J.C. appears to be founded on the idea that a judge may not have actually “relied on” a given “stereotype or improper inference”, as in cases such as Quartey (see J.C., at paras. 71-74). I would give effect to this question pursuant to the first question that I have described above. In my view, appellate courts must address the question of “reliance” at the outset of the analysis, rather than as part of a secondary “materiality” analysis after identifying an error. This encourages restraint and simplicity in terms of how appellate review is conducted by obviating the need for unnecessary scrutiny into the reasonableness of hypothetical generalized expectations or “stereotypes” that were not actually relied on by the trial judge.

[173]                     However, unlike the framework proposed in J.C., this question does not turn on whether the judge’s reliance on a generalized expectation was “material” or “important” to the analysis. Whenever an appellate court identifies the trial judge as having relied on a generalized expectation, the analysis proceeds to the second question that I describe below, regardless of whether that reliance was material or important. Again, an affirmative answer to this first question does not yet amount to an error.

[174]                     I come to this conclusion because, as I will explain, the question of whether an error was “material” to the conviction plays no role in identifying an error of law. Rather, it is only after an error of law has been identified on an appeal that the burden falls on the Crown to demonstrate whether the error was “material” to the conviction. The particulars of the Crown’s burden in this respect differ depending on the identity of the appealing party and the nature of the alleged error.

[175]                     In an appeal by the accused from a conviction, the question of materiality is addressed under the curative proviso in s. 686(1)(b)(iii) of the Criminal Code . Whenever an appellate court identifies an error of law by the trial judge in a conviction appeal, it may nonetheless decline to intervene under the curative proviso on the basis that no substantial wrong or miscarriage of justice has occurred. The curative proviso may be invoked in two circumstances: (1) where the error can be considered “harmless”, or (2) where the Crown’s case is “overwhelming” (R. v. Sarrazin, 2011 SCC 54, [2011] 3 S.C.R. 505, at para. 25 (emphasis deleted)). The burden falls on the Crown to demonstrate one of these requirements (R. v. Van, 2009 SCC 22, [2009] 1 S.C.R. 716, at para. 34). In either instance, “the underlying question is always whether the verdict would have been the same if the error had not been committed” (Van, at para. 36).

[176]                     The principles of appellate review for legal error and the application of the curative proviso must be kept distinct. The Criminal Code  and this Court’s jurisprudence mandate a structured analysis. The appellant must first demonstrate an error of law. Only then does the Crown, should it choose to rely on the curative proviso, bear the burden to satisfy one of the bases for its application. This division of responsibility “acknowledges that, when an error of law . . . is established, a presumption is created in favor of allowing the appeal, unless it is shown that the error or irregularity is curable” (R. v. Khan, 2001 SCC 86, [2001] 3 S.C.R. 823, at para. 88; see also Morrissey). The burden on the Crown is heavy, “reflecting the limited role of an appellate court and the need to safeguard the criminal justice process from the risk of wrongful convictions” (R. v. Samaniego, 2022 SCC 9, at para. 162, per Côté and Rowe JJ., dissenting). It would effectively reverse this burden to require the appellant in a conviction appeal to demonstrate materiality when identifying an error of law, as this simply mirrors the considerations under the curative proviso. For example, the Court of Appeal for Ontario in J.C. considered that a reversible error will arise only where “it cannot safely be said that the trial judge would have reached the same conclusion without the error” (para. 73). It is difficult to see how one could ever conclude that “the verdict would have been the same if the error had not been committed” under the curative proviso if it had already been decided that the judge may not have reached the same conclusion without the error.

[177]                     The Court of Appeal in J.C. may have mistakenly imported the requirement for materiality from two contexts in which materiality is required. However, there is a distinct rationale for the materiality requirement in those contexts, which cannot be translated to appeals from convictions on an error of law.

[178]                     First, the requirement for an error of law to be material arises in appeals by the Crown from acquittals (R. v. Graveline, 2006 SCC 16, [2006] 1 S.C.R. 609, at para. 14). In such appeals, the burden remains with the Crown to prove both the existence of legal error as well as the materiality of that error. The burden on the Crown is a “heavy one”, recognizing that acquittals ought not to be overturned lightly (R. v. Morin, [1988] 2 S.C.R. 345, at p. 374; Graveline, at para. 15; S. Coughlan, Criminal Procedure (4th ed. 2020), at p. 587). Thus, “different policy considerations apply in providing the Crown with a right of appeal against acquittals” (R. v. Biniaris, 2000 SCC 15, [2000] 1 S.C.R. 381, at para. 33). The same concerns about limiting Crown appeals cannot be imputed onto accused persons in appeals from convictions where, again, errors of law are presumed to have prejudiced the accused unless the Crown can prove otherwise under the curative proviso.

[179]                     Second, the issue of materiality arises in the context of misapprehensions of evidence amounting to a miscarriage of justice (see Morrissey; R. v. Lohrer, 2004 SCC 80, [2004] 3 S.C.R. 732; R. v. Sinclair, 2011 SCC 40, [2011] 3 S.C.R. 3). Here, the appellant in a conviction appeal must prove materiality because misapprehensions of evidence are, generally speaking, factual issues, not questions of law. There is no presumption that factual issues are material and thus constitute a miscarriage of justice, such that the grounds for appellate intervention under s. 686(1) (a) of the Criminal Code  are engaged. As explained by Justice Doherty in Morrissey, “[w]here the error is one of law the Crown bears the burden of demonstrating that the error did not result in a miscarriage of justice. Where the error is not one of law alone the appellant bears that burden” (p. 540).

[180]                     With this precision in hand, I turn to discuss the second question that an appellate court should ask, if it has determined that the trial judge relied on a generalized expectation.

B.            If the Trial Judge Relied on a Generalized Expectation, Was the Expectation Reasonable?

[181]                      Again, it is not an error of law per se for a trial judge to rely on a generalized expectation as a logical benchmark when assessing the evidence. Generalized expectations are necessarily relied on as inputs in the inference-drawing process, as well as when determining credibility. However, as I have explained, generalized expectations are based not on the evidence, but on common sense and human experience. Some generalized expectations may be widely held by the community but may be revealed over time to be inaccurate or unreliable, including through developments in psychological or social science research. Other expectations may not be widely held and may instead be the product of the trial judge’s own intuition, which may sometimes fail to accurately or reliably predict the behaviours or experiences of ordinary people. It is not controversial that inferences themselves cannot be based on “[i]ntuitive notions of probability at odds with objectively relevant probability theory” or “unsupportable stereotypical or prejudicial reasoning” (Hill, Tanovich and Strezos, at § 31:25). Similarly, where an inference is drawn pursuant to a generalized expectation, that expectation must reflect some measure of objective common sense rather than an individual’s assumption that may be intuitive but, ultimately, inaccurate. Thus, “if the trial process is a search for the truth, then misplaced assumptions about human behaviour which drive the trier of fact to draw incorrect inferences from the evidence must be unmasked if this process is not to be subverted rather than furthered” (Marquard, at pp. 265-66 (emphasis added), per L’Heureux-Dubé J., dissenting).

[182]                      I would therefore propose the following question for appellate review. If, under the first question, the appellate court concludes that the trial judge did rely on a generalized expectation in assessing the evidence, then the appellate court should determine whether the generalized expectation was reasonable. A standard of reasonableness imparts a measure of objectivity and community consensus in shaping the boundaries of a judge’s reliance on common sense and human experience to make decisions, thus ensuring that triers of fact do not rely on generalized expectations that are inaccurate or unreliable. It also recognizes that as society evolves, its understanding of what is reasonable may change.

[183]                      I note that the standard could have been couched in terms of “accurate”, “true”, “plausible”, etc. In my opinion, “reasonableness” is preferable. It strikes a balance between: (a) ensuring that triers of fact do not rely on unacceptably prejudicial or inaccurate generalizations on one hand, and (b) respect for the trier of fact’s role and the necessity of common sense and human experience in fact-finding on the other hand.

[184]                      Where a trial judge relies on an unreasonable generalized expectation in assessing the evidence, then the judge has committed an error of law. This will invite appellate intervention, subject to the Crown successfully invoking the curative proviso or, in an appeal by the Crown, subject to the Crown’s ability to demonstrate the materiality of the error (as explained above, these inquiries are distinct from the identification of legal error). I add that the fact that this is a “reasonableness” standard does not foreclose appellate review on a correctness standard for errors of law; other standards with an element of “reasonableness” (e.g., the existence of reasonable and probable grounds (R. v. Shepherd, 2009 SCC 35, [2009] 2 S.C.R. 527, at para. 20)) can entail a question of law.

[185]                     Similar to how drawing the line between a reasonable inference and speculation is “not always easy” (Villaroman, at paras. 38 and 43), it is difficult to set out, in the abstract, what amounts to a reasonable generalized expectation with perfect precision. It will be up to appellate courts to consider the reasonableness of a proposition derived from common sense or human experience. The crucial point, however, is that a generalized expectation must be a reasonably accurate reflection of what is true in most circumstances, such that it can be used as a reliable benchmark to assess the evidence.

[186]                     Examples of reasonable generalized expectations abound in the jurisprudence. Often, these generalized expectations relate to human experience or behaviour. This includes the “common-sense assumptions” I have referred to that a person foresees the natural consequences of their actions or is unlikely to make false statements against their own interest. Such expectations are frequently relied on in the fact-finding process because they are understood as reasonably accurate reflections of ordinary human behaviour or experience, until proven otherwise. This does not mean that these expectations are invariably true, only that they are usually true. The reasoning process depends on these “assumptions about the ordinary conduct of people in assessing credibility, and in assessing circumstantial evidence” (R. v. R.R., 2018 ABCA 287, 366 C.C.C. (3d) 293, at para. 5; see, e.g., A.R.D., at para. 100; see R. v. White, 2011 SCC 13, [2011] 1 S.C.R. 433, at para. 67). Thus, what is available as a generalized expectation in one case will be available in any other case (this relates to the principle of universality, which I discuss below).

[187]                     Where the boundaries of common sense or human experience end and the trial judge relies on an inaccurate proposition masquerading as “common sense” or “human experience”, the expectation cannot be said to be a reasonable reflection of what is true in most circumstances and should not be used as a benchmark to assess the evidence. Myths and stereotypes are recognized examples of inaccurate generalized expectations that “jeopardize the courts’ truth-finding function” (R. v. A.G., 2000 SCC 17, [2000] 1 S.C.R. 439, at para. 2). For example, the common law historically reflected the generalized expectation that victims of sexual assault are likely to report the acts, and that the failure to do so suggested that the acts did not occur. While at one time this was considered a fair expectation reflected in the “doctrine of recent complaint”, improved understanding over time has shown it to be an inaccurate and therefore unreasonable generalized expectation (see R. v. D.D., 2000 SCC 43, [2000] 2 S.C.R. 275, at paras. 60-63 and 65). However, myths and stereotypes are not exhaustive examples of unreasonable generalized expectations. As I will explain, my view is that an inaccurate generalized expectation about ordinary human behaviour should not be relied on to assess evidence, whether or not it can be characterized as a myth or stereotype.

[188]                     I stress that this inquiry targets a specific and limited part of the fact-finding process. Again, generalized expectations based on common sense and human experience are an input into the fact-finding process. They serve as a logical benchmark against which to assess the evidence to make conclusions of credibility or inferences of fact. Appellate review under this question is concerned solely with that benchmark, and not with the evidence being interpreted or any conclusions drawn from that evidence. Where a trial judge relies on a reasonable generalized expectation, an appellate court may not interfere with the judge’s assessment of the evidence based on that expectation just because it would have come to a different conclusion.

[189]                     The foregoing sets out my view of how appellate courts should assess the reasonableness of generalized expectations. In the following sections, I will discuss two further points arising from the Crown’s submissions: (1) whether the reasonableness of a generalized expectation is a question of law, and (2) whether the framework set out by the Court of Appeal for Ontario in J.C. should be endorsed, including the “rule against ungrounded common-sense assumptions” and the “rule against stereotypical inferences”.

(1)          Question of Law or Fact?

[190]                     Central to the parties’ disagreement on the legal principles engaged in these appeals is the issue of whether a trial judge’s reliance on a generalized expectation engages a question of law or a question of fact. This characterization is important to both appellate jurisdiction and the standard of appellate review.

[191]                     It is not disputed that there are limitations to reliance on generalized expectations that amount to questions of law. Notably, various expectations founded on myths or stereotypes about complainants, particularly women and children, in the context of sexual assault trials have been identified as errors of law, either by statute or by the courts. For example, s. 276(1)  of the Criminal Code  proscribes the drawing of inferences from the complainant’s past sexual activity to conclude on the question of consent or the credibility of the complainant — these are often referred to as the “twin myths”. Courts have identified other expectations that, as a matter of law, may not be relied on, such as the expectation that a person who has been sexually assaulted would necessarily exhibit avoidant behaviour towards the aggressor (see A.R.J.D.; see also, e.g., R. v. Mills, [1999] 3 S.C.R. 668, at para. 90; R. v. Osolin, [1993] 4 S.C.R. 595, at p. 670).

[192]                     To be clear, nothing in my reasons should be understood as altering or diluting the jurisprudence on recognized myths and stereotypes in the context of sexual assault trials. Such myths and stereotypes — including the “twin myths” set out in the Criminal Code  and other discredited expectations of how individuals, disproportionately women and children, behave during sexual encounters — have no place in the law and are appropriately denounced as errors of law (A.R.J.D., at para. 2). Nor should my reasons be understood as suggesting that all persons endure the same experiences of prejudice and marginalization, or to the same degree, irrespective of characteristics such as gender, age, and race.

[193]                     However, the Crown suggests that the categories of generalized expectations that may not, as a matter of law, be relied on are limited to what may be considered a discredited “myth” or a “stereotype”. I disagree. In my view, such a narrow approach creates artificial distinctions in the law, unduly limits the important role of appellate courts in the criminal justice system, and risks undermining public confidence in the administration of justice. Rather, for the reasons below, appellate review for legal error should extend to the reasonableness of any generalized expectation relied on by the trial judge.

[194]                     On a conceptual level, the underlying rationale for reliance on a generalized expectation is the same, regardless of whether the generalized expectation can be considered a “myth”, a “stereotype”, or something else. While the prevalence and experience of people subject to such generalized expectations will differ, all generalized expectations purport to derive from the same source: “common sense” and “human experience”. I am not persuaded that there is a principled basis to separate particular categories of generalized expectations for distinct treatment. Such an exercise invites artificial distinctions in the law with no clear boundaries.

[195]                     On a practical level, intermediate appellate courts have increasingly identified concerns about trial judges relying on unreasonable generalized expectations beyond the paradigmatic categories of myths and stereotypes about complainants in sexual assault cases. Professor Dufraimont observes that other categories have emerged in the jurisprudence, such as stereotypes about men and persons accused of sexual assault, or on grounds other than sex and gender such as against Indigenous and racialized people (see (2021), at pp. 546-50 and 557-59). In my view, these concerns cannot be disregarded out of hand as mere interventionism.

[196]                     I also agree with the intervener Association québécoise des avocats et avocates de la défense that the Crown’s narrower view would fail to account for generalized expectations that could not accurately be classified as “myths” or “stereotypes” about complainants in sexual assault cases yet would nonetheless be unacceptable to rely on in a criminal trial. For example, the trial judge in J.C., in rejecting the accused’s evidence, relied on the expectation that people (namely, people accused of sexual assault) generally do not achieve the “politically correct” ideal of obtaining consent for each escalating sexual act. In Roth, the trial judge appeared to assume that the accused, being a power lifter, was “somehow less susceptible to the effects of fatigue resulting from a long day, physical activity and the consumption of alcohol” (para. 70). These sorts of expectations are difficult to properly characterize as a “myth” or a “stereotype” as those terms are commonly understood in the jurisprudence (e.g., in Roth, the Court of Appeal for British Columbia considered that the trial judge did not rely on a “stereotyp[e]” about “male behaviour” (para. 66)), but I agree with the appellate courts in those cases that such expectations cannot reasonably be considered an accurate reflection of general human behaviour by which matters as significant as a criminal trial can be decided.

[197]                     The Crown asserts that appellate review of generalized expectations should be narrow because such expectations are deployed as part of the factual assessment of the evidence. Of course, deference is owed to trial judges on factual matters. However, not all issues that involve matters of evidence are necessarily questions of fact to which deference is owed. This Court has acknowledged at least four types of cases where the mishandling of evidence can constitute an error of law: (1) it is an error of law to make a finding of fact for which there is no evidence; (2) the legal effect of findings of fact or undisputed facts raises a question of law; (3) the assessment of the evidence based on a wrong legal principle is an error of law; and (4) the failure to consider all of the evidence in relation to the ultimate issue of guilt or innocence is an error of law (R. v. J.M.H., 2011 SCC 45, [2011] 3 S.C.R. 197). Thus, reliance on “myths” or “stereotypes” is recognized as an error of law despite relating to the assessment of evidence, as it amounts to an assessment of evidence based on a wrong legal principle — the third category set out in J.M.H. (see A.R.D., at para. 28; Coughlan, at p. 589, fn. 135). Again, in my view, the same logic extends to all reliance on unreasonable generalized expectations.

[198]                      I find support for my conclusion in the relevant policy reasons for why appellate courts defer to trial judges’ factual findings on one hand, and the wide discretion for appellate intervention on questions of law on the other hand.

[199]                     Appellate courts defer to trial judges on factual matters for at least three overarching policy reasons, as described in Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235 (a civil case): (1) to limit the number, length, and cost of appeals; (2) to promote the autonomy and integrity of trial proceedings; and (3) to recognize the expertise of the trial judge and the judge’s advantageous position to assess the evidence (paras. 16-18).

[200]                     In my view, reviewing unreasonable expectations on a standard of correctness would have a limited impact on these policy considerations. Concerns about preserving public resources or promoting the autonomy and integrity of trials, while important, are secondary to the rights of the accused in a criminal proceeding, in light of the interests at stake and Parliament’s decision to provide broad access to a first level of appeal (see, e.g., Sarrazin, at para. 27, affirming the high standard for the curative proviso). Rather, it is the recognition of the trial judge’s expertise and advantageous position that forms the central basis for appellate deference on factual matters in the context of criminal proceedings (see, e.g., R. v. Spencer, 2007 SCC 11, [2007] 1 S.C.R. 500, at para. 17; R. v. Beaudry, 2007 SCC 5, [2007] 1 S.C.R. 190, at para. 62). However, trial judges are not more experienced or in a more advantageous position than appellate judges in identifying a reasonable generalized expectation based on common sense or human experience. To the contrary, the basis for the general reliability of such expectations is that they are “common” or the shared experience of the entire community.

[201]                      In contrast, “the primary role of appellate courts is to delineate and refine legal rules and ensure their universal application” (Housen, at para. 9). Appellate courts therefore maintain a broad scope of review on questions of law. The principle of universality dictates that legal questions require clear and consistent answers in order to maintain public confidence in the administration of justice. Further, the public expects courts to fulfill a law-making function by bringing a measure of expertise on the art of just and practical rule-making. Appellate courts make law not only for the case under review, but also for future cases (ibid.).

[202]                     In the context of appeals from criminal matters, appellate courts play an especially important role in reviewing trial proceedings to prevent wrongful convictions and ensure that all convictions arise through a fair process that maintains public confidence in the administration of justice. As explained in Biniaris, “[c]riminal appeals on questions of law are based in part on the desire to ensure that criminal convictions are the product of error-free trials. Error-free trials are desirable as such, but even more so as a safeguard against wrongful convictions” (para. 26). It has been explained that “[a] person charged with the commission of a crime is entitled to a fair trial according to law. Any error which occurs at trial that deprives the accused of that entitlement is a miscarriage of justice” (Fanjoy v. The Queen, [1985] 2 S.C.R. 233, at p. 240; Morrissey, at p. 541). Any approach to appellate review that dilutes the important role of appellate courts would do significant damage to the criminal justice system as a whole.

[203]                     The characterization of an issue as a question of law is also significant to the grounds for appellate jurisdiction in criminal matters, especially the limited grounds for Crown appeals from acquittals. The Crown maintains a limited right to appeal on questions of law, which provides an avenue to balance the legitimate needs of a criminal justice system operating under the rule of law against the accused’s entitlement to finality when acquitted of criminal charges. Parliament and the courts have recognized the importance of protecting the dignity, privacy, equality, and other interests of all participants in the criminal justice system — including complainants — in order to maintain public confidence in the administration of justice (see, e.g., R. v. Ewanchuk, [1999] 1 S.C.R. 330). If the reasonableness of a generalized expectation were treated as a question of fact, then the Crown would not be able to appeal from acquittals that rest on generalized expectations that do not amount to “myths” or “stereotypes”, no matter how unreasonable or contrary to society’s collective expectations they are.

[204]                     In my view, these considerations support the application of a correctness standard when determining the reasonableness of a generalized expectation. This inquiry is not dependent on the evidence and instead rests on what is reasonably true in most instances. The need to preserve public confidence in the administration of justice mandates that verdicts in criminal cases — whether convictions or acquittals — not be founded on assumptions that are not reasonably accurate reflections of what is true in most circumstances. Appellate courts play an important role in ensuring the consistency and legitimacy of judicial reliance on generalized expectations, for the benefit of not only the case under review, but all future cases.

[205]                     Appellate courts should take caution that nothing in these reasons should be construed as meaning that the Crown can appeal an acquittal merely on a reassessment of the facts.

[206]                     We should not undermine the important institutional role of courts of appeal. Trial judges and courts of appeal have complementary roles, the latter being a safety net for errors by the former. In criminal law, effective appellate review is critical to avoiding improper verdicts. In earlier times, courts of appeal too readily intervened to substitute their own assessment of the evidence for that of the trial judge. More recently, under the banner of “deference”, there is increasing risk that the “safety net” role of courts of appeal is being unduly weakened.

[207]                     The palpable and overriding standard of review can be misused to generate a decision-making “black box” that facilitates ad hoc decision-making. By this, I mean an approach whereby if the Court of Appeal agrees with what the trial judge has done, it “shows deference”, while where the appellate court prefers another outcome, it labels the trial judge’s decision as a palpable and overriding error and substitutes its preferred outcome.

[208]                     Of course, this can also occur in applying the correctness standard, as was so in the “bad old days” where appellate courts too readily substituted their own assessment of the evidence. But, application of the palpable and overriding standard can be an opaque process. By contrast, the correctness standard, by its nature, requires setting out expressly an alternative line of reasoning and demonstrating why it should be followed. Properly applied, correctness demands greater conceptual clarity and analytical rigour. In part, this perspective underlies my analysis.

(2)          The “Rules” in J.C.

[209]                     Having set out my view of the scope and standard of appellate review, I turn to consider whether the Court of Appeal for Ontario’s approach in J.C. should be endorsed. As I have explained, J.C. represents a notable effort to address the concerns identified by intermediate appellate courts in many recent decisions. The Crown directs its attention to the “rule against ungrounded common-sense assumptions” described in J.C. and says that the rule should be rejected because it encourages appellate courts to interfere too readily with trial judges’ reliance on generalized expectations, as exemplified by the Court of Appeal for British Columbia in Mr. Tsang’s case.

[210]                     I would not adopt the articulation of the “rule against ungrounded common-sense assumptions” as described in J.C. It implies that a generalized expectation can be rejected merely because it is not “grounded” in the evidence. However, generalized expectations are not themselves grounded in the evidence, but are instead grounded in common sense and human experience. They nonetheless play a necessary role in the reasoning process when assessing the evidence.

[211]                     In my view, the difficulty with the standard set out in J.C. is exemplified by the Court of Appeal’s application of the proposed rule to the circumstances of that case. The accused testified about his practice of securing consent from the complainant on each progressive stage of sexual activity. The trial judge concluded that the accused’s testimony was “not in accord with common sense and experience about how sexual encounters unfold” (R. v. J.C., 2018 ONSC 5547, at para. 88 (CanLII)). The Court of Appeal considered that the trial judge came to this conclusion based on the assumption that “people engaged in sexual activity simply do not achieve the ‘politically correct’ ideal of expressly discussing consent to progressive sexual acts” (para. 97); in the Court of Appeal’s view, this ran afoul of the rule against ungrounded common-sense assumptions because the judge’s assumption was “a bald generalization about how people behave” and was “not derived from anything particular to the case, or any evidence before the trial judge on how all sexual encounters unfold” (para. 96). With respect, this is not a workable standard for appellate scrutiny. The lack of evidence supporting a generalized expectation cannot itself be a basis to reject reliance on that expectation. The Crown cannot be expected to elicit evidence on how sexual encounters ordinarily unfold in every sexual assault trial before a trial judge can rely on their common sense or human experience with respect to human sexual behaviour (see R.R., at para. 8).

[212]                     While I would not adopt J.C.’s articulation of a “rule against ungrounded common-sense assumptions”, this does not mean that the Court of Appeal in J.C. did not identify a legitimate concern with the trial judge’s reasoning in that case. The Court of Appeal concluded that the impugned expectation also contravened the “rule against stereotypical inferences”, as “it presupposes that no-one would be this careful about consent” (para. 97), and “the behaviour the trial judge rejected as ‘too perfect’, ‘too mechanical’, and ‘too politically correct’ to be believed is encouraged by the law, and certainly prudent” (para. 98). Respectfully, it is difficult to characterize the judge’s expectation in that case as a “stereotype” (To what group of persons is this stereotype directed?). However, I observe that the Court of Appeal in J.C. considered it an error of law under this rule to rely not only on stereotypes, but also on “erroneous common-sense assumptions” (para. 63). As I have explained, I would characterize all such assumptions within the broader category of unreasonable generalized expectations.

[213]                     Thus, the trial judge’s generalized expectation in J.C. was problematic not because it was ungrounded in the evidence, or because it was a “stereotype”, but because it is unreasonable to expect that people generally do not achieve the “politically correct” ideal of obtaining consent. The result and underlying rationale in J.C. are therefore compatible with my proposed framework, even though I would not adopt the same formulations of the principles as the Court of Appeal for Ontario.

C.            Did the Trial Judge Rely on a Generalized Expectation as Itself a Conclusive and Indisputable Fact?

[214]                     As I have explained, trial judges have considerable latitude to rely on reasonable generalized expectations as a logical benchmark in assessing the evidence. An appellate court may not interfere with the judge’s assessment of the evidence just because it would have come to a different conclusion from the evidence and any reasonable generalized expectations relating to that evidence.

[215]                     However, there is an important limit on the use of even a reasonable generalized expectation: The trial judge cannot treat the generalized expectation as itself a conclusive and indisputable fact, such that the judge ignores or forecloses their mind to the evidence. This is because people may always act contrary to a generalized expectation of what common sense or human experience would ordinarily anticipate. The trial judge’s duty is to determine on the evidence what really happened. For example, when drawing inferences from circumstantial evidence, a judge must consider the evidence “in light of all of the evidence and the absence of evidence, assessed logically, and in light of human experience and common sense” (Villaroman, at para. 30 (emphasis added)). Thus, generalized expectations based on human experience and common sense are only one consideration, which assist with interpreting the evidence. In the end, the focus must remain on the evidence.

[216]                     There is nothing novel about this aspect of my proposed framework. There are at least two reasons why the foregoing amounts to an error of law. First, it is an error of law for a trial judge to fail to consider all of the evidence on the ultimate issue of guilt or innocence — this is the fourth category of error described in J.M.H. (paras. 31-32; see also Walle, at para. 46). Second, where the judge treats a generalized expectation as itself an indisputable fact, what the judge is really doing is taking judicial notice of a fact. This is subject to the strict test for judicial notice — a standard that will rarely, if ever, be met by a generalized expectation about people due to the variability of human experience and behaviour. Where a factual conclusion is based neither on the evidence, nor on judicial notice, then it is speculation. This is an error of law, falling under the first category of errors described in J.M.H. — a finding of fact for which there is no evidence (para. 25; see Schuldt v. The Queen, [1985] 2 S.C.R. 592).

[217]                     For example, in L. (J.), the trial judge did not engage with the complainant’s and accused’s testimony and instead “relied on two facts to explain why he accepted that the complainant communicated a lack of consent to further sexual contact and that the [accused] pressed on in an attempt at intercourse”, one of which was “his finding that it ‘defies reason and common sense’ that ‘a young woman would go outside wearing a dress in mid-December, lie down in dirt, gravel and wet grass and engage in consensual sexual activity’” (para. 43). The Court of Appeal for Ontario considered that, although trial judges are permitted to exercise common sense when deciding a case, the judge erred “by relying on an assumption regarding what young women will and will not do, as if it were a fact” (para. 47 (emphasis added)). Coupled with the judge’s failure to consider the accused’s evidence or to explain why it failed to raise a reasonable doubt, this amounted to an error of law.

[218]                     Of course, a trial judge is not required to refer to every item of evidence or to explain how each piece of evidence was assessed (J.M.H., at paras. 31-32). “A trial judge must consider all of the evidence in relation to the ultimate issue but unless the reasons demonstrate that this was not done, the failure to record the fact of it having been done is not a proper basis for concluding that there was error in law in this respect” (R. v. Morin, [1992] 3 S.C.R. 286, at p. 296).

[219]                     In addition, this aspect of my proposed framework is of considerably less utility to the Crown on an appeal from an acquittal than an accused on an appeal from a conviction. This is for several reasons. First, an error of law due to the absence of evidence for a finding of fact “will happen as regards an acquittal only if there has been a transfer to the accused by law of the burden of proof of a given fact” (Schuldt, at p. 604). This is because “absent a shifting of the burden of proof upon the accused there is always some evidence upon which to make a finding of fact favourable to the accused” (p. 610). Relatedly, a trial judge’s decision to acquit based on a reasonable doubt “is not a finding of fact but instead a conclusion that the standard of persuasion beyond a reasonable doubt has not been met” (J.M.H., at para. 25). A reasonable doubt leading to an acquittal therefore does not need to arise from the evidence and can instead arise from the absence of evidence (paras. 26-27; Villaroman, at para. 36). Thus, when reviewing trial judges’ reasons for legal error, appellate courts must take particular care in Crown appeals from acquittals: “Caution must be taken to avoid seizing on perceived deficiencies in a trial judge’s reasons for acquittal to create a ground of ‘unreasonable acquittal’ which is not open to the court under the provisions of the Criminal Code  . . .” (R. v. Walker, 2008 SCC 34, [2008] 2 S.C.R. 245, at para. 2).

V.           Application to These Appeals

[220]                     Having set out what I consider to be the three questions to ask on appellate review of a trial judge’s use of generalized expectations, I turn to the purported errors in each of the two appeals before this Court.

A.           Mr. Kruk

[221]                     Mr. Kruk was charged with one count of sexual assault from an incident at his home in the early hours of May 27, 2017. Earlier that evening, he came upon the heavily intoxicated complainant on the street. The complainant testified that she woke up in Mr. Kruk’s bed with her pants off, Mr. Kruk on top of her, and his penis inside her vagina. Mr. Kruk testified that he took care of the complainant throughout the evening and intended to drive her to her home in the morning. He denied any sexual touching of the complainant and said the complainant removed her own pants while intoxicated. He proposed that the complainant woke up, still intoxicated, found her pants were off, and assumed the worst.

[222]                     The trial judge found Mr. Kruk’s testimony did not raise a reasonable doubt. In accepting the complainant’s testimony despite the reliability concerns from her intoxication, the judge stated: “She said she felt [the accused’s] penis inside her and she knew what she was feeling. In short, her tactile sense was engaged. It is extremely unlikely that a woman would be mistaken about that feeling” (para. 68 (CanLII)). The Court of Appeal concluded the judge erred in law by engaging in “speculative reasoning” and making an assumption on a matter that was not a matter of common sense or the proper subject of judicial notice.

[223]                     I disagree with the Court of Appeal. Answering the three questions I have described above leads me to the conclusion that the trial judge committed no error.

[224]                     Under the first question, the trial judge did rely on a generalized expectation about people — here, about physical perception. There was no evidence about whether this complainant was unlikely to be mistaken about the feeling of vaginal penetration. Thus, the judge assessed the plausibility of the complainant’s evidence with reference to an expectation of what the judge believed people — or, more specifically, women — in general would or would not experience.

[225]                     As I have explained, the fact that the judge relied on a generalized expectation is not itself problematic. To the contrary, it is a necessary part of judicial reasoning to assess evidence in relation to a benchmark of what might be ordinarily expected. However, the fact that the trial judge relied on a generalized expectation raises the possibility of an error of law under the second and third questions that I have described. My analysis therefore proceeds to those questions.

[226]                     Under the second question, I conclude that it is a reasonable generalized expectation that a woman is unlikely to be mistaken about the feeling of vaginal penetration. This is not to say that this is an incontrovertible fact about all women in all instances, or that a woman can never be mistaken. However, as a general proposition, it is perfectly reasonable. As the Crown observes, a sexual act of this nature would have a profound and traumatic impact on the bodily integrity of an individual, and ordinary people would not generally require special knowledge of such an aspect of human perception to assess this sort of evidence. To suggest otherwise would resurrect the long-rejected view that a complainant’s evidence about sexual assault cannot, without corroborating evidence, support a conviction. Further, it would impose an unnecessary, unrealistic, and invasive burden on the Crown to adduce expert medical evidence about the complainant’s basic physiology in every sexual assault case.

[227]                     The Court of Appeal implied that further detail from the complainant’s testimony about her experience may have assisted, noting that the complainant “was not asked, for example, to put into words what [the feeling of penetration] felt like, whether she experienced any pain, whether she had been injured or even why she felt so confident about her testimony” (para. 55 (CanLII)). I disagree that a more detailed description of the complainant’s physical experience would have assisted in this case. Nothing would have turned on such gratuitous evidence; the complainant’s testimony, if accepted, readily made out the actus reus of sexual assault.

[228]                     Mr. Kruk suggests that it was problematic for the trial judge, a man, to rely on a generalized expectation about what women would likely feel, as the judge would have “no personal experience regarding the matter” (R.F., at para. 56; see also para. 64). I disagree. While the trial judge would not have direct life experience of this precise experience, it is perfectly reasonable for any person to expect another person to not be mistaken about the sensation of having their bodily integrity violated in such a manner.

[229]                      Mr. Kruk also says that the reasonableness of an expectation depends on the particular circumstances of the case, and that “[t]he further the circumstance comes from the norm to the outlying extreme, the less likely that it is a matter of common sense” (transcript, at p. 70). Thus, in his submission, the fact that the complainant “was extremely intoxicated and disoriented” meant that the judge’s generalized expectation was no longer reasonable (ibid.). This misunderstands the nature of a generalized expectation and its role in the reasoning process of fact-finding. It is sufficient for an expectation to be reasonable in a generalized manner, meaning it would be true in most circumstances. The complainant’s state of intoxication and the resulting concerns about her reliability were certainly important aspects of the evidence that the judge was required to consider when assessing the complainant’s evidence, but the particulars of the evidence do not alter the reasonableness of the generalized expectation at issue. Rather, the complainant’s evidence needed to be assessed both with reference to a reasonable benchmark, and in light of the evidence as a whole, including the evidence of intoxication. This is also the case, for example, with consideration of the “common sense inference” that a person intends the reasonable and probable consequences of their actions, even where there is evidence that the accused was impaired (Walle, at paras. 58-67). This relates to the third question of my proposed framework, to which I now turn.

[230]                     In my view, the Court of Appeal’s conclusion most closely reflects a concern arising under the third question. The Court of Appeal considered the trial judge erred by making a “finding” that was not the “proper subject of judicial notice” (para. 67), which caused the trial judge to “engag[e] in speculative reasoning that was not grounded in the evidence” (para. 68). If this were true, I agree that this would amount to an error of law. However, I disagree that the trial judge committed this error. The trial judge did not treat the generalized expectation that a woman is unlikely to be mistaken about the feeling of vaginal penetration as a conclusive and indisputable fact in this case (i.e., as a fact drawn on judicial notice). Rather, he properly relied on that generalized expectation as a benchmark for assessing the plausibility of the complainant’s evidence.

[231]                     Crucially, the judge did not foreclose his mind to the rest of the evidence on the basis that a woman could never be mistaken about the feeling of vaginal penetration. The judge relied on circumstantial evidence that he considered was “consistent with a sexual encounter having occurred between the accused and the complainant”, including their respective states of undress and the removal of the complainant’s pants (para. 69). The judge also considered “the failure of the accused to contact the complainant’s parents in the overall circumstances . . . to be consistent with an intention to take sexual advantage of a young woman he knew to be in an extremely compromised and vulnerable state” (para. 70). The judge was also clearly alive to the reliability concerns about the complainant’s evidence (para. 48). The judge was entitled to find the complainant’s evidence reliable on this point by assessing that evidence with reference to a reasonable generalized expectation about human perception as a benchmark, and in light of the other circumstantial evidence.

[232]                     For the reasons above, I conclude that the trial judge did not commit the error alleged by Mr. Kruk or identified by the Court of Appeal.

B.            Mr. Tsang

[233]                     Mr. Tsang was charged with one count of sexual assault from an incident in his car in the early hours of December 29, 2018. He and the complainant met at a music concert and attended a subsequent event at the Commodore Ballroom in Vancouver. The two then departed in Mr. Tsang’s car. The complainant testified that on the way to drop her off, Mr. Tsang stopped the car at a parking lot, where he asked to make out in the back seat. She consented, expecting no more. However, without her consent, Mr. Tsang forced her to perform oral sex and penetrated her vaginally and anally. He then dropped her off at her friend’s home and abruptly drove away as soon as she exited the vehicle. Evidence from a sexual assault medical examiner noted trauma to the genital area, including a tear to the vagina. Mr. Tsang testified that the complainant suggested and led all instances of sexual activity, including the oral sex and digital vaginal penetration; that she asked to be “spanked” and for other instances of “rough sex”; that he never penetrated her anally; and that they did not engage in penile-vaginal intercourse because he could not find a condom. The trial judge rejected Mr. Tsang’s evidence and concluded that Mr. Tsang controlled the complainant throughout the night and “orchestrated” the events. She found the complainant to be credible, accepted her evidence, and convicted Mr. Tsang (2020 BCPC 306).

[234]                     The Court of Appeal concluded that the judge erred in rejecting Mr. Tsang’s testimony by making three “assumptions” or “generalizations” about human behaviour that amounted to “speculative reasoning”: that (1) a person would not ask to be spanked while engaging in sexual foreplay “out of the blue”; (2) a controlling person would not refrain from engaging in vaginal intercourse because of the absence of a condom; and (3) a person would not abruptly and unceremoniously leave another with whom he had engaged in consensual sex (paras. 73-74).

[235]                     I will address each of these three purported errors in turn.

(1)          A Person Would Not Ask To Be Spanked While Engaging in Sexual Foreplay “Out of the Blue”

[236]                     Under the first question that I have described, I conclude that the trial judge did not rely on a generalized expectation of human behaviour when she found that the complainant did not ask to be spanked or otherwise encourage “rough sex”. I note that while the Court of Appeal at times focused on the judge’s rejection of Mr. Tsang’s evidence that the complainant “asked him to spank her” (para. 28), it expressed a broader concern that the judge “rejected [Mr. Tsang’s] evidence that the complainant encouraged rough sex” (para. 43).

[237]                     The trial judge’s rejection of Mr. Tsang’s “rough sex” evidence was not based on a generalized expectation about whether any person would ask to be spanked or otherwise encourage rough sex. Instead, it was based on the judge’s assessment of the evidence about the particular individuals in question and her conclusions about Mr. Tsang’s credibility in particular. The judge’s reasons clearly indicate that she did not believe Mr. Tsang’s story that the complainant had instigated all flirtatious or sexual advances that evening. The judge had, at this point in her reasons, noted that the complainant “had expressed no interest” in Mr. Tsang throughout the night until the incident in his vehicle, and that she “had been with another man earlier in the evening” (para. 121 (CanLII)). The judge had also rejected Mr. Tsang’s testimony about the complainant initiating a kiss on the Commodore Ballroom dance floor — a conclusion that the Court of Appeal acknowledged was open to the judge based on the evidence (para. 39). Further, and importantly, the trial judge considered Mr. Tsang’s evidence about rough sex was contrived to explain the presence of injuries (para. 158). Taken at its highest, Mr. Tsang’s evidence suggested the complainant wanted to be “spanked” or to have “harder” digital penetration. The judge considered, in light of the complainant’s injuries, that this sort of purportedly consensual “rough sex” failed to explain the severity of the injuries actually exhibited by the complainant. It was in light of these conclusions that the judge found it was “not believable” that this complainant would have asked to be spanked or was “gearing up for rough sex” with Mr. Tsang (para. 126).

[238]                     With respect, the Court of Appeal reweighed the evidence. It downplayed aspects of the judge’s reasoning by concluding, for example, that “little turns upon the events that preceded those that took place in the parking lot” (para. 42) and that it could not be said that other aspects of the evidence “weighed heavily” in the judge’s assessment of the evidence (para. 45). The trial judge’s assessment of the evidence was clearly informed by the entire transaction of events during the evening and her overall assessment of the evidence. The Court of Appeal was not entitled to substitute its own view of how much “turned” on different parts of the narrative.

[239]                     Accordingly, the trial judge did not rely on a generalized expectation of what people are likely to do during sexual activity. There is therefore no need to consider the second and third questions that I have described in relation to this alleged error.

(2)          A Controlling Person Would Not Refrain From Engaging in Vaginal Intercourse Because of the Absence of a Condom

[240]                     As with the first purported error, I conclude under the first question that the trial judge did not rely on a generalized expectation about human behaviour when she rejected Mr. Tsang’s evidence that he refrained from vaginal intercourse because of the absence of a condom. There is therefore no need to consider the second and third questions that I have described.

[241]                     The impugned statement in the trial judge’s reasons is as follows:

The accused’s description about the prospect of intercourse being thwarted by the lack of a condom in his car when there was one available was contrived in my view and contrary to the level of control he conveyed about that evening and in court. He said [the complainant] told him she wanted to give him oral sex to finish him and she did so again. He said he used his hand to grab her hair lightly and guide her. He recalled that she choked on it a bit and then used her mouth and hands again. Mr. Tsang said he asked [the complainant] if she spat or swallowed. When [the complainant] answered “swallow; spit is for losers”, he ejaculated inside her mouth. I find Mr. Tsang’s behaviour to have been focussed on himself at this time but for which it would have been apparent that [the complainant] was only participating so that he would finish and not because she wanted to or was enjoying it. [Emphasis added; para. 129.]

[242]                     The Court of Appeal considered that the trial judge relied on a “prejudicial stereotype” about “controlling” people to assume that such people would not abstain from sex because of the lack of a condom (para. 65). In my view, there is no generalized expectation about “controlling” people at play in the trial judge’s reasons. The judge concluded that Mr. Tsang himself was “controlling” and had “orchestrated” events throughout the night, based on circumstances such as Mr. Tsang’s insistence that the complainant and her friend leave their fanny packs and belongings in his car and his encouragement of further drinking. The trial judge thus found it contrived that the inability to find a condom would have “thwarted” his motivation to have sexual intercourse with the complainant, in light of all of the other evidence of his behaviour throughout the evening. Her overall conclusion was, of course, informed by the complainant’s direct evidence that he did instigate sexual intercourse, which the judge accepted. This conclusion was also informed by the judge’s finding that Mr. Tsang had exhibited few inhibitions through the night and was focussed on his own pleasure while in the car, as indicated by his indifference to hearing the complainant crying and gagging during oral sex.

[243]                     The Court of Appeal again reweighed the evidence in order to identify an error. It did not question the trial judge’s conclusion that Mr. Tsang was “controlling” and had “orchestrated” the events throughout the night. It acknowledged the various aspects of the evidence that the judge considered. Despite acknowledging these points, the Court of Appeal downplayed the significance of the trial judge’s impressions of Mr. Tsang, again by stating that “little turns upon the events that preceded those that took place in the parking lot” (para. 42). There was no basis for the Court of Appeal to intervene with the trial judge’s appreciation of the evidence as a whole.

[244]                     As I would dismiss this argument under the first question, there is no need to consider the reasonableness of any generalized expectation under the second question, as none was relied on. However, I would add that, in my view, the Court of Appeal’s characterization of the trial judge’s analysis as relying on a “prejudicial stereotype” about “controlling people” distorts a proper understanding of stereotypes. “Controlling people” are not members of a particular group who are subject to biases that could fairly be described as “stereotypes”. The Court of Appeal’s approach would imply that any characterization about a person’s personality, such as being “aggressive” or “careless”, could be labelled a “stereotype”, even where such characterizations arise from the evidence of the person’s behaviour.

(3)          A Person Would Not Abruptly and Unceremoniously Leave Another With Whom He Had Engaged in Consensual Sex

[245]                     I turn to the final error identified by the Court of Appeal. The impugned passages of the trial judge’s reasons are as follows:

[Mr. Tsang’s] lack of interest . . . at [the complainant’s] invitation to meet again was at odds with his evidence that they . . . just had a great time, but is consistent with a non-consensual event where he got what he wanted without regard for her and drove away.

 

. . .

 

Mr. Tsang’s testimony aligned with that of [the complainant] and [her friend] that he drove off as soon as [the complainant] got out of his car and he did not watch her go inside the house. I find this fact more consistent with [the complainant’s] claim of non-consensual sex than with Mr. Tsang’s version of what had just happened. He took off right away because of what he had just done to her and because she meant nothing to him. It is inconsistent with his evidence that he wanted to pleasure [the complainant] that night. [Emphasis added; paras. 131 and 153.]

[246]                     Under the first question that I have described — and unlike the first two purported errors — I consider that the trial judge did rely on a generalized expectation about how people ordinarily behave after a consensual sexual encounter. I disagree with the Crown that the judge was solely considering the consistency between Mr. Tsang’s evidence that he and the complainant had just had a great time with the evidence of Mr. Tsang’s abrupt driving away. The judge did not simply say that Mr. Tsang’s evidence was inconsistent, but she affirmatively considered that Mr. Tsang driving away was “consistent with a non-consensual event where he got what he wanted without regard for her” and “more consistent with [the complainant’s] claim of non-consensual sex”. There was nothing in the evidence to suggest that Mr. Tsang was the type of person who would or would not abruptly drive away from a person with whom he had engaged in consensual sex. Rather, the judge considered that most people do not do this.

[247]                     Furthermore, under the second question, I consider that it is unreasonable to expect any logical connection between an individual waiting for their sexual partner to enter a home and the consensual or non-consensual nature of the preceding encounter. Driving away could certainly indicate impoliteness or a lack of social tact. Yet this is a far cry from suggesting that the person has just participated in a non-consensual sexual encounter. This cannot be understood to be a reasonably accurate reflection of what is true in most circumstances, such that it can be used as a reliable benchmark to assess the evidence.

[248]                     Despite this error of law, I would uphold the conviction under the curative proviso on the basis that no substantial wrong or miscarriage of justice has occurred. In my view, this error of law was “so harmless or minor that it could not have had any impact on the verdict” (Van, at para. 34). As the Crown observes, this generalized expectation played a very minor role in the trial judge’s analysis. The judge’s verdict clearly turned on her favourable assessment of the complainant’s credibility, as well as her complete rejection of Mr. Tsang’s testimony as inconsistent with the events throughout the evening and contrived to explain away the complainant’s injuries. While the judge did consider Mr. Tsang’s abrupt driving away as further evidence of a non-consensual encounter, she only entertained this evidence in response to Mr. Tsang’s suggestion that he and the complainant had just had a great evening. In these circumstances, it is inconceivable that the trial judge would have had a reasonable doubt about consent had she not relied on the evidence about his abrupt driving away.

VI.         Conclusion

[249]                     For the reasons above, I would allow both appeals and restore the convictions.

 

APPENDIX

 

 

 

                    Appeals allowed.

                    Solicitor for the appellant: Ministry of Attorney General — Criminal Appeals, Vancouver.

                    Solicitors for the respondent Christopher James Kruk: Johnson Doyle Nelson & Anderson, Vancouver.

                    Solicitors for the respondent Edwin Tsang: Fowler & Blok, Vancouver.

                    Solicitor for the intervener the Attorney General of Alberta: Alberta Crown Prosecution Service — Appeals and Specialized Prosecutions Office, Calgary.

                    Solicitors for the intervener the Independent Criminal Defence Advocacy Society: Thorsteinssons, Vancouver; Pringle Law, Vancouver.

                    Solicitors for the intervener the Criminal Lawyers’ Association (Ontario): Gorham Vandebeek, Toronto.

                    Solicitors for the intervener the Trial Lawyers Association of British Columbia: Peck and Company, Vancouver; Olthuis van Ert, Vancouver.

                    Solicitors for the intervener Association québécoise des avocats et avocates de la défense: Hugo Caissy (ad hoc), Amqui, Que.; Beaudry Roussin, Québec.

                    Solicitors for the interveners the West Coast Legal Education and Action Fund Association and the Women’s Legal Education and Action Fund Inc.: Megan Stephens Law, Toronto; West Coast LEAF, Vancouver; Women’s Legal Education and Action Fund (LEAF), Toronto.