No More Duncan Instruction in Canada (so why say it?)

Duncan instruction

The Duncan instruction, originating from the 1981 decision of the English Court of Appeal in R v Duncan, 73 Cr App R 359, is a trial judge’s direction given to a jury where a mixed statement – containing both inculpatory and exculpatory parts – is admitted into evidence:

[W]here appropriate, as it usually will be, the judge may, and should, point out that the incriminating parts are likely to be true (otherwise why say them?), whereas the excuses do not have the same weight.

This instruction was approved by the House of Lords as recently as 1995. However, in two decisions released last Friday, the Supreme Court of Canada (“SCC”) has ruled that the Duncan instruction should not be adopted by the Canadian courts.

R v Rojas

In R v Rojas, [2008] 3 SCR 111 [Rojas], two brothers, Hugo and Miguel Rojas, were convicted of second degree murder. The evidence against them was largely circumstantial, except for a number of incriminating statements they had made to David Miranda who, by burning the car containing the body of the victim, became their accomplice. Two grounds of appeal were advanced before the Court: firstly, that the Vetrovec v The Queen, [1982] 1 SCR 811 warning with respect to Miranda’s testimony resulted in allowing the jury to make impermissible use of his evidence of the brothers’ statements, and secondly, that it was improper for the trial judge to give the jury the Duncan instruction. The SCC unanimously dismissed the appeal; with regard to the Duncan instruction, Justice Charron, for the Court, held that:

[I]t is dangerous for the judge to instruct the jury in a manner that suggests that inculpatory and exculpatory statements ought to be weighed differently. Such ‘common sense comments are better left to the advocacy of counsel. Therefore, I conclude that the Duncan instruction should not be adopted by Canadian trial courts (para 40).

R v Illes

In R v Illes, [2008] 3 SCR 134 [Illes], the companion case to Rojas, there were again two grounds of appeal, one of which was with respect to the trial judge’s Duncan instruction to the jury. Having decided in Rojas that it would be improper to use the Duncan instruction in Canadian courts, the question before the Court was whether the error was such to warrant a new trial. In a 4-3 split, the majority, per Justices Fish and LeBel, found that it was, while the dissent, per Justice Charron, would have dismissed the appeal.

Comparing Rojas and Illes

Although the Court rejected the use of the Duncan instruction, it also indicated that its use would not necessarily constitute a reversible error. Rather, it would be “a matter to be determined on a case-by-case basis, having regard to the charge as a whole and the evidence at trial” (Rojas, para 5). In particular, Justice Charron noted in para 41 of Rojas that, “where the instruction is couched in terms of a legal presumption, the prejudicial effect may be difficult to overcome.”

Thus, the difference between the two cases was the actual words used by the trial judges in giving the Duncan instruction. In Rojas, the judge had instructed the jury as follows:

Since some of the alleged statements of each accused contain exculpatory statements or denials which tend to exonerate them, you must consider that, in the ordinary course of human behaviour, it often occurs that statements of an incriminating nature such as admissions or confessions are likely to be true, otherwise why say them. On the other hand, excuses for one’s own behaviour do not necessarily carry the same persuasive weight (para 27).

In Illes, the trial judge told the jury:

On the other hand, any portion of the statements that tend to incriminate Mr. Illes in the commission of the offence only assist the Crown’s case if you are convinced that portion is true beyond a reasonable doubt. This is because the law presumes any incriminating part of the accused’s statement is likely to be true, otherwise why would an accused say so (para 15). [emphasis added]

Surprisingly, despite the fact that the trial judge in Illes had “couched [the jury instruction] in terms of a legal presumption” in the way that she had warned against in Rojas, Justice Charron nevertheless found that:

[H]aving regard to the contents of the letters in question, their admitted fabrication by Mr. Illes, the respective use made of the evidence at trial by the Crown and the defence, and the instructions to the jury as a whole, I agree with the majority in the Court of Appeal that this unfortunate instruction, although given in error, could not realistically have impacted on the verdict (Illes, para 59).

In contrast, the majority found that the jury instruction could be interpreted in a way that it might have encouraged the jury to discredit the Defence theory and place more weight on the Crown’s interpretation of the mixed statements. Notably, what appears to have convinced the majority is the length of time and the apparent difficulty the jury had in reaching its verdict:

It is nonetheless worth noting that the jury deliberated over four days; that the jurors indicated in a note to the judge on the evening of the first full day of deliberations that they were split on their decision; and that the jury was re-charged with a portion of the instructions containing the impugned language on the very day that it finally rendered the verdict (para 22).

A Question regarding Illes

What is confusing about the majority decision in Illes is that it seems to imply the trier of fact should take the exculpatory parts of mixed statements at face value, rather than within the context of the other evidence advanced in trial. As Justices Fish and LeBel indicated in para. 9:

[T]he letters urge the accused’s associates ‘not to confess to anything you haven’t done;’ state that ‘NOTHING EVER HAPPENED, so ANYTHING ANYBODY MAKES UP IS A LIE;’ accuse the police of using ‘KGB police trick[s];’ and warn that the police ‘will try EVERYTHING to try to build a case, to try to fabricate evidence’ (emphasis in original). While many of these statements are implausible in the context of the other evidence at trial, they are all exculpatory statements on their face. [emphasis added]

Also, in para. 19:

The potential for jury confusion on the scope of the Duncan instruction is even more apparent when one appreciates that a single statement can be exculpatory on its face, but inculpatory when regarded in light of other evidence. Thus, for example, the claim that ‘NOTHING EVER HAPPENED, so ANYTHING ANYBODY MAKES UP IS A LIE’ is clearly exculpatory if accepted as literally true. But when read with other statements in the same letter, such as speculation that Ms. Ovalle may have killed the victim or that the victim was in hiding, the statement might appear disingenuous. When one considers that the statement was addressed to Mr. Madinsky, the only other person who could have killed the victim, the claim that ‘NOTHING EVER HAPPENED’ takes on a decidedly conspiratorial valence. [emphasis added]

Is the not jury supposed to decide on a verdict based on the totality of the evidence? If the statements that appear to be exculpatory on its face become inculpatory when considered in its context, is it not likely that the jury would have inevitably reached the same guilty verdict?


The rejection of the Duncan instruction would appear to be consistent with Canadian jurisprudence. Prior to Rojas and Illes, courts in Canada had been divided on the validity and usefulness of the Duncan instruction, and the problem exacerbated by the inconsistent use of different wording of the instruction. As noted earlier, it would seem “dangerous” or improper “for the judge to instruct the jury in a manner that suggests that inculpatory and exculpatory statements ought to be weighed differently,” and eliminating its use would prevent confusion and inconsistent verdicts in future cases.

You may also like...

Join the conversation

Loading Facebook Comments ...