La juge Karakatsanis — À l’instar du juge Leurer, dissident, nous sommes d’avis qu’une directive spécifique de type Hibbert (R. c. Hibbert, 2002 CSC 39,  2 R.C.S. 445) était requise dans les circonstances de la présente affaire. L’appel est accueilli, essentiellement pour les motifs exposés par le juge Leurer. La déclaration de culpabilité est annulée et un nouveau procès est ordonné.
 The “danger of wrongful conviction arising from faulty but apparently persuasive eyewitness identification has been well documented” (R v Hibbert, 2002 SCC 39 at para 51,  2 SCR 445 [Hibbert], referring to Peter deCarteret Cory, The Inquiry Regarding Thomas Sophonow: The Investigation, Prosecution and Consideration of Entitlement to Compensation (Winnipeg: Manitoba Justice, 2001) at 31–34). Justice Doherty has stated that the “spectre of erroneous convictions based on honest and convincing, but mistaken, eyewitness identification haunts the criminal law” (R v Quercia(1990), 1990 CanLII 2595 (ON CA), 60 CCC (3d) 380 at 383 (Ont CA)). David Tanovich, Louis Strezos and S. Casey Hill, McWilliams’ Canadian Criminal Evidence, loose‑leaf (Rel 2021‑No 1) 5th ed (Toronto: Thomson Reuters, 2013) at para 32:10 fn 6, say, with reference to international jurisprudence, that “[s]imilar cautions are universally accepted”.
An instruction that such identification should be accorded “little weight” was found not to go far enough to displace the danger that the jury could still give it weight that it does not deserve.
 Hibbert provides direction to courts on how to approach the type of eyewitness evidence that lies at the heart of this appeal – what is often referred to as “in-dock” identification testimony. More specifically, the case has led to a requirement that jurors be given “what is known as a Hibbert instruction”, going beyond standard cautions, in situations where there is suspect in-court identification testimony (R v Jack, 2013 ONCA 80 at para 31, 294 CCC (3d) 163 [Jack]). This Court summarized the effect of Hibbert in R v Bigsky, 2006 SKCA 145, 217 CCC (3d) 441 [Bigsky]:
 The Supreme Court of Canada in Hibbert found that the trial judge should have cautioned the jury more strongly that the identification of the accused in court was highly problematic as direct reliable identification of the perpetrator of the offence. The Court noted the danger associated with eyewitness in-court identification is that it is deceptively credible, largely because it is honest and sincere. An instruction that such identification should be accorded “little weight” was found not to go far enough to displace the danger that the jury could still give it weight that it does not deserve. Moreover, it should also have been stressed that the impact of the victim having seen the accused arrested by the police as her alleged assailant could not be undone. And this was so even though the trial judge had addressed the frailties of the identification evidence. The Supreme Court of Canada allowed the appeal and ordered a new trial.
 Since Hibbert, appeal courts have emphasized that a caution – going beyond a general one relating to the frailties of eyewitness identification evidence, and instead specifically directed to instructing jurors as to the dangers of placing reliance on in-courtidentification testimony – is required in cases where the in-court identification is suspect. In R v Pelletier, 2012 ONCA 566 at para 93, 291 CCC (3d) 279, Watt J.A. stated, with reference to many authorities dating from before and after Hibbert, that “as a general rule, in-dock identifications are entitled to little weight in the assessment of the adequacy of the prosecution’s proof on the issue of identity”. I will give several additional examples of cases where general instructions relating to the frailties of eyewitness identification evidence were found to be insufficient and jury charges were held to be deficient because jurors were not adequately cautioned about the dangers of in-court identification testimony.
The value, if any, of an in-court identification is inherently tied to what supports that testimony.
 Second, I would not want these reasons to be read as suggesting that an in-dock identification of an accused person is in most cases impermissible. To the contrary, it is often required. In Hibbert, Arbour J. said that, in that case, the in-court identification “served to confirm that the accused was, in the opinion of Ms. McLeod and Ms. Baker, the same man they saw throughout the chain of events (from arrest through to the second trial)”. For that reason, “[i]n that sense, despite its almost total absence of value as reliable positive identification, the evidence of the witnesses may be given some weight at least for that purpose” (at para 49, emphasis added). Indeed, the law recognizes that in-court identification may be required as a precondition to the admission of prior identification evidence: R v Tat (1997), 1997 CanLII 2234 (ON CA), 117 CCC (3d) 481 at paras 40 and 46–50 (Ont CA). However, as Doherty J.A. explained in Tat, the value, if any, of an in-court identification is inherently tied to what supports that testimony:
 Clearly, the evidence of the prior descriptions given and the prior identifications made by the identifying witness constitute prior consistent statements made by that witness. Generally speaking, evidence that a witness made prior consistent statements is excluded as irrelevant and self-serving. However, where identification evidence is involved, it is the in-court identification of the accused which has little or no probative value standing alone. The probative force of identification evidence is best measured by a consideration of the entire identification process which culminates with an in-court identification: e.g. R. v. Langille, supra, at 555; Di Carlo v. United States,6 F.2d 364 (2d cir. 1925) at 369, per Hough J., concurring; Clemons v. United States, 408 F.2d 1230 (D.C. cir. 1968) at 1243. The central importance of the pre-trial identification process in the assessment of the weight to be given to identification evidence is apparent upon a review of cases which have considered the reasonableness of verdicts based upon identification evidence: e.g. see R. v. Miaponoose (1996), 1996 CanLII 1268 (ON CA), 110 C.C.C. (3d) 445 (Ont. C.A.).
 If a witness identifies an accused at trial, evidence of previous identifications made and descriptions given is admissible to allow the trier of fact to make an informed determination of the probative value of the purported identification. The trier of fact will consider the entirety of the identification process as revealed by the evidence before deciding what weight should be given to the identification made by the identifying witness. Evidence of the circumstances surrounding any prior identifications and the details of prior descriptions given will be central to that assessment.
(Emphasis added, footnote omitted)
Juries should receive a caution about the dangers of an in-court identification when the first actual identification of an accused occurs in court. Beyond this, consideration of when a special caution is required, and what it must contain, begins with an understanding as to what prior identification a witness has made of the accused before testifying. Account must also be taken of the quality or basis for such identification. Finally, attention must be paid to the opportunities that exist for that identification to change over time, and why that might be the case.
 I find it to be neither necessary nor appropriate to attempt to construct an exhaustive list of criteria as to when a Hibbert-type instruction is required. However, several principles emerge from the case law that I have reviewed and guide me in this case. In this regard, juries should receive a caution about the dangers of an in-court identification when the first actual identification of an accused occurs in court. Beyond this, consideration of when a special caution is required, and what it must contain, begins with an understanding as to what prior identification a witness has made of the accused before testifying. Account must also be taken of the quality or basis for such identification. Finally, attention must be paid to the opportunities that exist for that identification to change over time, and why that might be the case. As is evident from the case law I have reviewed, the reasons why these are paramount considerations are that the evidentiary value of the in-court identification is a direct product of these three factors and, conversely, the dangers associated with an in-court identification increase as the quality of any pre-testimony identifications decreases.
 I begin with the simple, but important, observation that the judge did not give the jury the specific cautions about the evidence of Mr. Williams and Ms. Holmgren that, I conclude, the circumstances here demanded. The judge did not tell the jury, in the strongest terms, that it was required to approach Mr. Williams’s in-court identification evidence with extreme caution and explain to it why this is the case. The judge did notforcefully tell the jury that it could give virtually no weight to Mr. Williams’s in‑court identification of Mr. Clark as the man he saw assault Mr. Durocher. The judge also did notcaution the jury at all about Ms. Holmgren’s in-dock identification or instruct the jury that it would be very dangerous to attribute to Ms. Holmgren’s in-court identification of Mr. Clark any degree of certainty greater than what she had communicated to police on October 2, 2016.