The criminal justice system can be daunting, but you don’t need to go through it alone. Our Criminal lawyers are here to guide you every step of the way.
Contact Our Firm
Joseph A. Neuberger, Diana Davison, and Michael P. Bury, Neuberger & Partners LLP
Few kinds of evidence pose a harder problem for a criminal court than a memory that its owner had lost and then found. In R. v. B.B., 2023 ONSC 396, Justice J. Dawe of the Ontario Superior Court of Justice was asked to convict a father of sexually assaulting his daughter on the strength of a single recollection that, by the complainant’s own account, had been absent from her conscious mind for more than two decades before it returned as a sudden and vivid “flashback.” The judgment that followed is a careful, and ultimately cautionary, study in what the criminal standard of proof demands when the entire case rests on the reliability of one person’s recovered memory.
The court acquitted. But the reasoning behind the acquittal is what makes the decision worth reading closely, because it turns not on any finding that the complainant was dishonest, but on the far more uncomfortable question of whether a sincerely held memory can be trusted at all.
01 • THE ALLEGATION
The complainant, referred to throughout the judgment as T.B. to comply with a publication ban, was 34 years old when, over the 2018 Canada Day weekend, an ordinary gesture from her fiancé triggered what she described as a full and immediate recollection. She remembered lying in her parents’ bed as a child, her father reaching under her shirt and touching her nipples. She testified that she had no continuous memory of the event; rather, it had surfaced all at once, felt entirely real, and left her certain it had happened. On the basis of this recollection, her father, B.B., was charged with a single count of sexual assault said to have occurred when she was around eleven years old, in or about 1996.
B.B., 63 years old at trial and with no criminal record, testified in his own defence and denied that he had ever touched his daughter in the manner alleged.
Crucially, there was no other evidence. As the trial judge put it, this was a two-witness case: everything before the court came from T.B. and B.B., in the form of their testimony and a series of letters and text messages they exchanged in 2020. There were, in the court’s words, no contemporaneous or independent documents and no expert testimony on the science of memory from either side.
02 • THE LEGAL FRAMEWORK
The elements of the offence were not seriously in dispute. Sexual assault requires proof of unwanted sexual touching (the actus reus) together with the intention to touch (the mens rea). Because the alleged conduct occurred when the complainant was well below the then-applicable age of consent, and because the accused was her father, in a position of trust and authority, there was no question about lack of valid consent if the sexual touching took place as described.
That left a single real question, and the judgment is admirably direct about it: was the court satisfied beyond a reasonable doubt that the touching happened as T.B. now remembered it, given that she conceded she had no memory of it for over twenty years?
Justice Dawe grounded the analysis in orthodox principles. Proof beyond a reasonable doubt sits much closer to absolute certainty than to a mere balance of probabilities. Demeanour is a notoriously unreliable guide to accuracy, and the court expressly declined to place weight on how either witness presented. And credibility, the judgment notes, is best tested against consistency with independent and contemporaneous evidence — of which, here, there was none.
03 • THE HEART OF THE CASE: ASSESSING A RECOVERED MEMORY
The most valuable part of the decision is its frank treatment of the difficulty a judge faces when asked to evaluate recovered-memory evidence without expert assistance.
Justice Dawe refused both of the easy paths. It would be an error, he held, to discount T.B.’s account simply because she was an adult describing childhood events, and equally an error to accept it uncritically. He accepted, for the purposes of analysis, that some recovered memories may be genuine, and that T.B. might truly be recalling a real event. But he also accepted that her recollection might be a confabulation — a vivid, sincerely believed memory of something that never occurred.
What the judgment does especially well is expose the trap in relying on “common sense” here. Common sense, the court observed, is built out of a judge’s own necessarily limited experience. A judge who has never had a flashback of a buried trauma may wrongly assume such memories cannot be reliable; a judge who has never held a vivid but false memory may find it hard to believe memory can deceive so completely. On this terrain, intuition is as likely to mislead as to guide. Justice Dawe drew support from the extrajudicial writing of the late Justice Marc Rosenberg, who warned that repressed memory challenges the ordinary assumption that recollection fades with time, and that the very richness of a recovered memory can create a false impression of reliability.
Several specific features of T.B.’s evidence were examined and found wanting as indicators of reliability:
Drawing an analogy to the well-established disconnect between an eyewitness’s confidence and the accuracy of an identification, the court held that subjective certainty does not make a memory objectively reliable — particularly where that certainty had been reinforced by therapy that pre-supposed the memory’s authenticity.
There was no evidence before the court that a confabulated memory cannot be as richly detailed as a genuine one. Peripheral details, such as the layout of the parents’ bedroom, were things T.B. would have known in any event and could readily have been woven into an imagined event.
She struggled to place significant childhood milestones and misremembered relatively recent events, which made the exceptional vividness of the disputed recollection stand out rather than reassure.
While the court accepted that T.B.’s memory was not created by therapeutic suggestion — the flashback preceded the counselling — it could not exclude the possibility that details had been added or solidified over four years of revisiting and discussing the memory.
None of these observations amounted to a finding that T.B. was lying. The court was explicit that she was a credible and sincere witness. The concern was reliability, not honesty — a distinction the judgment keeps carefully in view throughout.
04 • THE CROWN’S KEY ARGUMENT: THE FATHER’S OWN WORDS
With no independent evidence available, the Crown’s principal submission was that B.B.’s own communications with his daughter in 2020 — although exculpatory on their face — should be read as confirming the reliability of her memory. In his messages and a letter, B.B. repeatedly said he did not remember the incident and insisted he would never have intentionally hurt her, rather than issuing a flat, categorical denial. An innocent man, the Crown contended, would simply have said “this never happened.” The qualified nature of the denials, on this theory, were said to betray a guilty mind.
Justice Dawe rejected the argument on several grounds.
First, he saw no genuine contradiction between denying an event and saying one does not remember it. A person speaking to a non-event can only ever speak from present memory; witnesses who refuse to concede that something they cannot recall might have happened are often accused of being rigidly dogmatic.
Second, he invoked the rule against stereotypical inferences articulated in R. v. J.C., 2021 ONCA 131. Just as it is an error to measure a complainant’s conduct against assumptions about how a “real” victim behaves, it is equally an error — and, the court suggested, a graver one — to measure an accused against a stereotype of how an innocent person “ought” to react, and then to recast facially exculpatory words as though they were confessions.
Third, and factually decisive, the court was not satisfied that B.B. even understood the full nature of the allegation when he wrote to his daughter. On the evidence, he may have believed only that she was accusing him of putting his hand up her shirt, not of touching her breasts or nipples. If so, his emphasis on the absence of any sexual intent — and his openness to the possibility of accidental contact — becomes entirely comprehensible.
Fourth, the messages had to be read in context. T.B. had raised the allegation in the immediate aftermath of a bitter family argument, against a backdrop in which she repeatedly framed any challenge to her account as “disrespect” and a “deal breaker” for their relationship. In her first letter she had, in effect, offered her father safe passage if he did not contest the reality of the memory. It was hardly surprising, the court found, that he accepted that invitation rather than mounting a frontal assault on what she called “her truth.”
Finally, and most importantly, the court identified the deeper flaw: the Crown’s argument threatened to reverse the burden of proof. Even if the court disbelieved B.B.’s denials entirely, disbelief of an accused’s account is not affirmative proof of guilt. Rejecting an exculpatory statement cannot be used as a “makeweight” to bolster otherwise insufficient prosecution evidence and bridge the gap in the Crown’s case.
05 • THE W.(D.) ANALYSIS AND THE VERDICT
Justice Dawe concluded by working through the familiar three-part framework from R. v. W.(D.), [1991] 1 S.C.R. 742, though he took the steps in reverse. He was not satisfied that T.B.’s recovered memory was reliable, and there was no independent evidence capable of dispelling the possibility that it was a confabulation — meaning the Crown’s case fell short even before the accused’s evidence was weighed. He could not reject B.B.’s testimony as incredible, though neither did he affirmatively believe it. This was not a case where the complainant’s evidence was so compelling that the defence account had to be rejected despite the absence of obvious flaws.
The court’s candour in its final passages is striking. Justice Dawe acknowledged that T.B.’s memory might be true, and that B.B. might be lying or genuinely forgetful. But he found it equally possible that B.B.’s denial was both honest and accurate, and that T.B.’s recollection was a sincerely believed figment of her imagination. He could not choose between these possibilities even on a balance of probabilities, let alone to the criminal standard. The result followed inexorably: a reasonable doubt, and an acquittal.
06 • WHY THE DECISION MATTERS
R. v. B.B. is not a pronouncement that recovered memories are inherently unworthy of belief. Justice Dawe was careful to leave open the possibility that such memories can be genuine, and he grounded his caution in the burden of proof rather than in any general skepticism about complainants. The decision’s importance lies elsewhere.
It is, first, a disciplined illustration of the difference between honesty and reliability. A witness can be entirely sincere and entirely mistaken, and the criminal law’s demand for proof beyond a reasonable doubt is indifferent to good faith. Where a conviction depends on the accuracy of a single, uncorroborated, long-dormant recollection, sincerity cannot do the work that proof requires.
Second, the judgment is a useful warning against the seductions of “common sense” in domains where ordinary experience gives out. On questions of how memory stores, loses, and reconstructs events, an untutored intuition is a poor instrument, and the decision is honest about the limits of a judge’s own knowledge in the absence of expert evidence.
Third, the treatment of the accused’s ambiguous statements is a meaningful application of the rule against stereotypical inferences to the defence side of the ledger. Courts have increasingly recognised that assumptions about how a “real victim” behaves are impermissible; R. v. B.B. confirms that assumptions about how an innocent person behaves are no more acceptable, and that reinterpreting cautious words as veiled admissions risks reversing the onus of proof.
For practitioners, the practical lesson is plain. In a recovered-memory prosecution without corroboration, the evidentiary gap is unlikely to be bridged by parsing an accused’s guarded reactions. And for anyone interested in how the presumption of innocence operates at its outer edge, the decision is a clear and thoughtful example of a court declining to convict not because it believed the accused, but because it could not honestly say it disbelieved him and the evidence of the complainant did not meet the threshold for reliability.
This commentary discusses a published decision of the Ontario Superior Court of Justice and is offered for general information only. It is not legal advice. A publication ban under s. 486.4(1) of the Criminal Code protects the identity of the complainant; the initials used in the judgment and reproduced here comply with that order. Neuberger & Partners LLP, Sex Assault Defence Lawyers Toronto.
