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
A look at R. v. T.L., 2026 ONCA 190
Joseph A. Neuberger, Michael P. Bury, and Diana Davison, Neuberger & Partners LLP, Toronto Criminal Lawyers
When several complainants describe similar assaults by the same person, that similarity can be powerful proof of guilt when used appropriately by the Crown. But what if they talked to each other first? At some point, comparing notes stops being corroboration and starts being contamination. The Ontario Court of Appeal’s recent decision in R. v. T.L. is a sharp reminder that a trial judge cannot have it both ways: a judge who treats the risk of tainted evidence as serious enough to exclude evidence for one purpose has to explain why that same risk disappears when it comes time to convict.
01 • THE FACTS
The appellant was convicted of sexual assault of two complainants, C.E. and S.S., at their high school. He, the two complainants, and a third woman, N.D., were all enrolled in a special education program for students with cognitive issues, and all had some form of intellectual disability.
The three women were friends. Each said the appellant had assaulted her in a small storeroom, or “cubby,” at the school. Crucially, they had discussed the assaults with one another before going to the police. N.D. testified that the three of them had met, agreed the appellant was a “problem,” and decided together to report him. She accepted the defence’s characterization that they were “pissed off,” had agreed to “get” the appellant, and wanted to “stick it to him.”
The evidence was also messy in the way real memory often is. S.S. initially reported an assault outdoors, and police found nothing on surveillance and closed the file; a later interview relocated the assault to the cubby. By the time of trial, both C.E. and S.S. said they could remember very little of the events themselves. Their answers about who told whom, and whether anyone had asked them to fabricate anything, shifted from question to question.
02 • TWO DIFFERENT RULINGS FROM ONE JUDGE
Here is where the case gets interesting. The trial judge had to make two separate decisions, and they pointed in opposite directions.
The Crown asked to use N.D.’s uncharged allegations and each complainant’s account across counts to help prove the appellant’s guilt. Similar fact evidence is presumptively inadmissible; the Crown must show its value outweighs its prejudice, and part of that analysis is ruling out the possibility that the accounts line up because the witnesses influenced each other. The judge found no deliberate collusion, but he could not rule out inadvertent tainting — the risk that friends with acknowledged memory difficulties had unconsciously shaped each other’s recollections through repeated conversation. Because the Crown had not disproven that possibility, the judge dismissed the application.
The same judge then accepted C.E. and S.S. as completely credible and reliable, rejected the appellant’s evidence, and convicted. On the tainting issue, he now stated that he found “no evidence of collusion or tainting of evidence as between CE and SS” — without explaining how the concern that had just sunk the similar fact application had evaporated.
03 • WHY THAT WAS A REVERSIBLE ERROR
The Court of Appeal allowed the appeal and ordered a new trial. The problem was not that the judge reached inconsistent results — a judge can dismiss a similar fact application and still convict. The problem was the absence of an explanation.
Tainting doesn’t have to be deliberate to matter. As the court reiterated, hearing another witness’s account can colour and reshape a person’s own description of events, whether anyone intends it. A trial judge faced with a live possibility of tainting must confront it directly and may rely on the evidence only if demonstrably satisfied the tainting did not affect the witness’s credibility or reliability.
The judge’s own similar fact ruling established that there was an “air of reality” to inadvertent tainting. Having found that, he could not simply assert the opposite at the conviction stage. He had to say why the risk that was fatal to admissibility of similar fact evidence no longer troubled him when assessing whether the complainants were telling the truth — including why he wasn’t worried that N.D. had tainted C.E. and S.S. Both sides had agreed N.D.’s “get him” evidence was relevant to that very question, yet the judge’s credibility analysis never mentioned it.
The Crown leaned on R. v. O.M., where a judge similarly dismissed a similar fact application over collusion concerns and later convicted. But O.M. actually cut the other way: there, the judge tackled collusion “head on” and explained why it didn’t undermine the witnesses. Here, there was no such reasoning to review.
04 • NO RESCUE FROM THE PROVISO
The Crown’s fallback was the curative proviso, which lets an appellate court uphold a conviction despite a legal error where the error is harmless. The court rejected it. The proviso generally has no place in a case that turns entirely on credibility and reliability, and this error was anything but trivial. There was significant evidence of at least inadvertent tainting that the judge ignored — and the risk was heightened, not lowered, by the witnesses’ circumstances. The complainants were suggestible, and by trial they had little independent memory of the events, which made their accounts more vulnerable to being shaped by what others had told them.
05 • THE TAKEAWAY
R. v. T.L. is a useful marker for anyone working with multiple-complainant cases. A finding that the Crown hasn’t disproven tainting isn’t a self-contained ruling that can be quarantined to the admissibility stage. Once a judge accepts there is a real possibility that witnesses influenced one another, that possibility follows the evidence into the credibility analysis, and it must be reckoned with out loud. Judges retain the discretion to believe witnesses who have spoken to each other — but they must show their work, and that obligation only grows where suggestibility and fragile memory are in play.
READ THE FULL DECISION HERE: R. v. T.L., 2026 ONCA 190
This post discusses a decision subject to a publication ban under s. 486.4 of the Criminal Code; the parties are referred to by initials as in the reported judgment. It is a general summary for information only and is not legal advice. Neuberger & Partners LLP, Sex Assault Defence Lawyers Toronto.
