“V was charged with historical sexual offences against a single complainant and tried before a judge and jury. The jury convicted him of sexual interference and invitation to sexual touching. The same jury acquitted him of sexual assault based on the same evidence. V appealed his convictions, asserting that they were inconsistent with his sexual assault acquittal and therefore unreasonable. The Crown cross‑appealed V’s acquittal, maintaining that the charge to the jury was so unnecessarily confusing that it amounted to an error in law and that the apparent inconsistency in the verdicts could be explained by the erroneous jury instructions, such that the guilty verdicts could not be considered unreasonable.
A majority of the Court of Appeal held that there was no legal error in the jury instructions and that the convictions for sexual interference and invitation to sexual touching were unreasonable, as they were inconsistent with the acquittal on the sexual assault charge. The majority quashed V’s convictions and substituted verdicts of acquittal, and upheld the acquittal on the sexual assault charge. The minority found legal error in the jury instructions and would have ordered a new trial on all three charges.”
The SCC (7:2, in part) allowed the appeal.
Justice Moldaver wrote as follows (at paras. 5-7, 28-36, 41, 74-75):
“…the Crown can seek to reconcile apparently inconsistent verdicts by showing, to a high degree of certainty, that the acquittal was the product of a legal error in the jury instructions, that the legal error did not impact the conviction, and that the error reconciles the inconsistency by showing that the jury did not find the accused both guilty and not guilty of the same conduct. If the Crown discharges its burden, appellate intervention on the conviction is not warranted because the verdicts are not actually inconsistent and thus not unreasonable on the basis of inconsistency.
…I am respectfully of the view that the trial judge misdirected the jury on the charge of sexual assault by leaving the jury with the mistaken impression that the element of “force” required for sexual assault was different than the element of “touching” required for sexual interference and invitation to sexual touching. In particular, the failure to instruct the jury in clear terms that the “force” required to establish sexual assault was one and the same as the “touching” required to establish the other two offences constituted non‑direction amounting to misdirection. The effect of this error on the apparently inconsistent verdicts is significant. A review of the charge to the jury as a whole enables me to conclude, with a high degree of certainty, that the error was material to the acquittal. Equally, I am satisfied that the error did not impact on the convictions; rather, it reconciles the apparent inconsistency in the verdicts. Accordingly, the verdicts are not actually inconsistent and the convictions are not unreasonable on the basis of inconsistency.
In the result, I would allow the appeal and restore the convictions. I would also set aside the acquittal and, as I will explain, enter a stay of proceedings on the charge of sexual assault.
The Criminal Code does not expressly identify inconsistent verdicts as a ground for setting aside a conviction. For an appellate court to interfere with a conviction on the ground that it is inconsistent with an acquittal, the court must find that the guilty verdict is unreasonable (R. v. Pittiman, 2006 SCC 9,  1 S.C.R. 381, at para. 6, citing Criminal Code, s. 686(1) (a)(i)). The accused bears the burden of establishing that a verdict is unreasonable (Pittiman, at para. 6).
In an appeal involving inconsistent verdicts, the applicable test to determine whether a verdict of a jury is unreasonable is: “Are the verdicts irreconcilable such that no reasonable jury, properly instructed, could possibly have rendered them on the evidence?” (Pittiman, at para. 10). Put another way, a conviction is unreasonable and must be set aside where the verdicts cannot be reconciled on any rational or logical basis and no properly instructed jury, acting reasonably, could have rendered the verdicts it did based on the evidence (R. v. McShannock (1980), 55 C.C.C. (2d) 53 (Ont. C.A.), at p. 56; Pittiman, at paras. 6‑7).
When verdicts cannot be reconciled and a jury that was properly instructed returns a conviction that is not supportable on the evidence presented at trial, the only available inference is that the jury acted unreasonably in arriving at the conviction (R. v. Biniaris, 2000 SCC 15,  1 S.C.R. 381, at para. 39). The jury may have reached a compromised verdict, misunderstood the evidence, or nullified by choosing to not apply the law — any of those paths to inconsistent verdicts reflects unreasonableness. In such cases, the conviction itself is unreasonable and appellate intervention is warranted.
The ultimate inquiry for appellate courts then is whether the verdicts are actually inconsistent. Apparently inconsistent verdicts can be reconciled on the basis that the offences themselves are “temporally distinct, or are qualitatively different, or dependent on the credibility of different complainants or witnesses” (Pittiman, at para. 8). If verdicts are reconciled to reveal a theory on which the jury could have returned the verdicts without acting unreasonably, the verdicts are consistent and appellate intervention is not warranted.
In my view, there are also cases, such as the one at hand, where the Crown can reconcile apparently inconsistent verdicts on the basis that they were the result of a legal error in the jury instructions. For such cases, I propose the following approach.
Where the Crown attempts to rebut an apparent inconsistency on the basis of a legal error, the burden shifts from the accused to the Crown. That burden is heavy. The Crown must satisfy the court to a high degree of certainty that there was a legal error in the jury instructions and that the error:
- (1) had a material bearing on the acquittal;
- (2) was immaterial to the conviction; and
- (3) reconciles the inconsistency by showing that the jury did not find the accused both guilty and not guilty of the same conduct.
If the court can find that these elements are satisfied with a high degree of certainty, the verdicts are not actually inconsistent. Instead, the legal error caused the jury to convict the accused either on different evidence or a different element than it believed was necessary for the charge on which it acquitted the accused. Any apparent inconsistency in the verdicts is thus reconciled, as the jury did not find the accused both guilty and not guilty of the same conduct. It follows that the jury did not act unreasonably in rendering their verdicts.
In assessing whether the Crown has satisfied its burden, the court must not engage in improper speculation about what the jury did and did not do. The appellate court must be able to retrace the reasoning of the jury with a sufficiently high degree of certainty to exclude all other reasonable explanations for how the jury rendered its verdicts. If it can, any concern about speculation falls away.
This approach respects the ordinary deference afforded to the presumed reasonableness of the jury by asking “whether the [apparently inconsistent verdicts] are supportable on any theory of the evidence consistent with the legal instructions given by the trial judge” (Pittiman, at para. 7). Where the conviction is supported by the evidence, as is always required, and the verdicts are not actually inconsistent, the jury’s entering of a conviction against the accused is not unreasonable and the conviction appeal should be dismissed. A jury does not act improperly by relying on a trial judge’s legal error. Put another way, the appellate court simply concludes that the jury acted reasonably based on the evidence and instructions before it. The conviction is thus reasonable and appellate intervention is not warranted.
On finding that a legal error shows that apparently inconsistent verdicts are not actually inconsistent, the appropriate remedy depends on whether the Crown has cross‑appealed the acquittal.
Under s. 686(8) of the Criminal Code , a court of appeal has the power whenever it exercises “any of the powers conferred by subsection (2), (4), (6) or (7) . . . [to] make any order, in addition, that justice requires”. The Criminal Code also vests that power in this Court (s. 695(1)). For an appellate court to issue an order under its s. 686(8) residual power, three requirements must be met (R. v. Thomas,  3 S.C.R. 535). First, the court must have exercised one of the triggering powers conferred under s. 686(2) , (4) , (6) or (7) . Second, the order issued must be ancillary to the triggering power. Consistent with the provision’s “broad remedial purpose”, this Court has taken a flexible approach in determining whether the order is “in addition” to the exercise of the triggering power (R. v. Hinse,  4 S.C.R. 597, at para. 30). In particular, the additional order need not directly advance the exercise of the triggering power (Hinse, at paras. 31‑32; see, e.g., R. v. Power,  1 S.C.R. 601, at pp. 615‑18; R. v. Provo,  2 S.C.R. 3, at pp. 19‑21; Terlecki v. The Queen,  2 S.C.R. 483, at pp. 483‑84). It is enough that the ancillary order not be “at direct variance with the court’s underlying judgment” (Thomas, at para. 17; see also R. v. Warsing,  3 S.C.R. 579, at paras. 72‑74). Third and finally, the order must be one that “justice requires”.
Here, the three requirements justify ordering a stay of proceedings on the sexual assault charge. First, the Court’s s. 686(8) residual jurisdiction is triggered by allowing the Crown’s appeal and setting aside the acquittal under s. 686(4) (b). Section 686(8) provides for residual jurisdiction where the appellate court “exercises any of the powers conferred by subsection . . . (4)”. Allowing an appeal and setting aside the verdict of acquittal constitutes one of those powers. It therefore triggers s. 686(8) jurisdiction even without a further order of new trial (R. v. Bellusci, 2012 SCC 44,  2 S.C.R. 509, at para. 39; R. v. Smith, 2004 SCC 14,  1 S.C.R. 385, at para. 22).”
Justice Brown (with whom Kasirer J. concurred) (dissenting in part) wrote as follows (at paras. 81-92):
“That the verdicts in this case are inconsistent is not in dispute. What divides us is what to do about it ⸺ or, more precisely, what a court can and cannot do about it. Our jurisprudence and the Criminal Code, R.S.C. 1985, c. C-46 , compel an inescapable, if unfortunate, result: a new trial on all three charges. My colleagues find such a result unpalatable ⸺ so unpalatable, that they seek to escape the otherwise inescapable by effectively overruling aspects of the relevant jurisprudence (notably, R. v. J.F., 2008 SCC 60,  3 S.C.R. 215), and eliding the limits of the Code ⸺ indeed, re‑writing (or, more accurately, re‑legislating) one of the Code’s limits on curial jurisdiction.
Specifically, my colleagues restore the respondent’s convictions on the basis that the verdicts are not inconsistent in the minds of the jury, and are therefore reasonable. But this ignores that the inconsistency of the verdicts is the very reason this appeal is before us. Again, there is no dispute that the verdicts are inconsistent. And yet, the fact of the inconsistency sends my colleagues off on a search to discover why the inconsistency exists. This is possible, they say, because what matters is not the actual legal elements of the offences themselves, but instead how they were likely understood in the minds of the jury ⸺ as can best be guessed by the reviewing court. And having pasted together a plausible explanation for why the jury reached inconsistent verdicts, it follows (ex hypothesi) that what appear to be inconsistent verdicts are no longer inconsistent: they are reconcilable, by pointing to misdirection.
This, however, ignores the appeals scheme of the Criminal Code and the law as it stands, which make clear that it simply does not matter why the inconsistency exists, or whether it can be explained away with a “high degree of certainty” (majority reasons, at para. 33). To explain, the Code codifies the common law rule of jury secrecy and provides asymmetrical rights of appeal to a convicted person and the Crown (s. 649 and R. v. Pan, 2001 SCC 42,  2 S.C.R. 344; ss. 675 and 676 ). These rules inform the remedies available on appeal, which depend on whether an appeal is from a conviction or an acquittal, and whether an appeal is from the verdict of a judge alone or a jury (ss. 686(2) to (4) ). To succeed on an appeal from acquittal, the Crown must establish legal error that “might reasonably be thought, in the concrete reality of the case at hand, to have had a material bearing” on the verdict (R. v. Graveline, 2006 SCC 16,  1 S.C.R. 609, at para. 14). And, of significance here, where the Crown successfully appeals from a verdict of acquittal by a jury, the only available remedy is a new trial (s. 686(4) (b)(i)).
It is plain from this scheme that Parliament’s crafting of the appeals provisions in the Code was governed by a fundamental truth about jury trials: we can never know for certain how a jury reached its verdict. A reviewing court can, of course, examine the verdict and the record, including the evidence, the arguments of counsel and the instructions of the trial judge. But a reviewing court cannot inquire into the jury’s reasons for arriving at the verdict (Pan, at para. 46). This is why, where the Crown successfully establishes legal error that meets the Graveline threshold, a conviction cannot be substituted for the acquittal. Again, Parliament, in specifying a new trial as the sole remedy, proceeded on the basis that ⸺ unlike in a judge alone trial ⸺ a reviewing court can only speculate about why the jury acquitted the accused. As a majority of this Court said in R. v. Morin,  2 S.C.R. 345, at p. 374, it cannot “predict with certainty what happened in the jury room”.
The significance of all of this is as follows: my colleagues think they know the precise effect and extent of the trial judge’s legal error in misdirecting the jury. But while they may think they know this (even to the “high degree of certainty” that they claim), and while I may think that my colleagues’ suppositions may well have merit, none of this matters. In crafting the Code’s provisions governing appeals from jury verdicts, Parliament did not care about the degrees of certainty at the reviewing court. What Parliament thought significant is that the absence of reasons for judgment by a jury means a reviewing court can never be certain what was in the minds of the jury. My colleagues undermine that clear legislative intent.
My colleagues’ line of reasoning also does violence to J.F. They dispute this, saying that J.F. (and, for that matter, R. v. Pittiman, 2006 SCC 9,  1 S.C.R. 381) should not be read as precluding the review of jury instructions in assessing the consistency of verdicts. But that is not the issue here, and indeed nobody suggests such a reading. In Pittiman and J.F., proper instruction is said to be presumed for the purposes of determining whether the verdicts are actually legally inconsistent. This is the basis on which a reviewing court can determine, as here, that a conviction for sexual interference and an acquittal for sexual assault arising out of the same delict are fundamentally inconsistent.
But my colleagues also say ⸺ and here they do run afoul of J.F., their protestations to the contrary notwithstanding ⸺ that J.F. does not preclude an inquiry into why a jury returned inconsistent verdicts, pointing to the absence of a clear error in the jury instructions in that case as explaining why the Court could not embark upon the same inquiry that my colleagues do here. This is a plainly erroneous reading of Fish J.’s reasoning for the Court in that case, where he wrote that “[i]n any event, as a matter of legal process and the legitimacy of verdicts, I would decline to uphold the respondent’s conviction on the ground that it can be reconciled with his acquittal on another count of the same indictment on the basis of a legal error at trial” (para. 21 (underlining added)). The insertion of the phrase “in any event” could not be clearer: the existence of a legal error does not reconcile inconsistent verdicts. That is an aspect of J.F. which my colleagues, albeit without acknowledging as much, effectively overrule.
The novelty of the framework now developed by my colleagues is revealed in their analogy to the reasoning of Paciocco J.A. in R. v. Plein, 2018 ONCA 748, 365 C.C.C. (3d) 437, wherein he held that “close examination of the trial judge’s reasons for judgment and court record [may provide a] rational or logical basis that can reconcile the verdicts” (para. 28 (emphasis added)). But, as his reasons indicate, that was a case in which a judge, sitting alone, rendered inconsistent verdicts. At the risk of stating the obvious, this is inapplicable to juries. But my colleagues are undeterred: “. . . so long as the reviewing court is able to retrace the jury’s reasoning to the high degree of certainty required”, the appellate review is akin to considering the reasons of a trial judge (para. 39 (emphasis added)). But retracing a jury’s reasoning, irrespective of the “degree of certainty”, is a type of review that: (1) Parliament has precluded; (2) this Court has never sanctioned; and (3) is, as a practical matter, impossible. On that last point ⸺ practical impossibility ⸺ my colleagues’ framework requires a reviewing court to be able to “exclude all other reasonable explanations for how the jury rendered its verdicts” (para. 35). But of course, that will never be possible. The best they can do is guess. And here, the possibility that the jury could have nullified or compromised (concluding, for example, that two convictions for approximately the same delict were sufficient, and three excessive) is, on my colleagues’ guess, unreasonable, and thus ruled out.
My colleagues respond by stating that commonplace appellate matters “requir[e] a reviewing court to consider . . . whether [jury members] might reasonably have changed their minds if the trial had unfolded differently” (para. 37), and point to (inter alia) the Graveline threshold of whether an error might reasonably be thought to have affected the result. This is, of course, true. But that proposition is not what divides us here. Pointing to instances that require reviewing courts to consider whether an error or omission might reasonably be thought to have affected the result (and, if so, whether a new trial should follow) is a distraction. My colleagues, with respect, are doing something quite different. Specifically, they say that, despite convictions and an acquittal for the same delict, and despite the existence of a legal error at trial, it is possible for them to conclude that the jury certainly thought the respondent is guilty. There is a world of difference between an appellate court saying (1) “this error, viewed objectively, may have made a difference to the jury”, and (2) “I know what this jury, viewed subjectively, was thinking”. The latter is unprecedented. Indeed, were this simply the “routine” matter that my colleagues say it is (at para. 38), they would not have to invent a new framework ⸺ let alone a new framework that breaks from this Court’s precedent ⸺ to do it.
My colleagues go further still to avoid being taken as having overruled J.F. on this point, insisting that nothing in their analysis “suggests that an error of law in the instructions to the jury necessarily makes improper verdicts proper or inconsistent verdicts consistent” ⸺ which would, of course, be contradictory to J.F. ⸺ because “[i]f the error can be isolated to the acquittal, it is not the error itself that reconciles the verdicts, but rather the further determination that the error did not affect the conviction” (para. 40). But there is simply no meaningful difference between (1) finding that an error reconciles inconsistent verdicts, and (2) finding them reconciled by the “determination that the error did not affect the conviction”. Under their framework of analysis, an error is found, the jury’s reasoning is (somehow) carefully retraced and reconstructed, and a determination is made that ⸺ in the jury’s minds ⸺ the offences were different. The error reconciles the inconsistent verdicts.
My colleagues conclude as much here, finding that in this case “the legal error reconciles the apparent inconsistency” (para. 70). This holding is manifestly inconsistent with J.F., where Fish J. wrote that “[i]mproper instructions do not make . . . inconsistent verdicts consistent” (para. 23). It is also at odds with Pittiman, where Charron J. offered the following routes to reconcile potentially inconsistent verdicts in the case of a single accused charged with multiple offences: the offences may be “temporally distinct, or . . . qualitatively different, or dependent on the credibility of different complainants or witnesses” (para. 8). I note the omission in Charron J.’s reasons of any reference to considering jury instructions; indeed, it could not have occurred to her to do so, since (until now) verdicts like those in the case at bar were considered actually inconsistent, not potentially or apparently inconsistent.
In restoring the respondent’s convictions, my colleagues say that this will be a rare result. I disagree. The result they reach is an invitation to routine speculation into the reasoning process of the jury, something (again) that the Criminal Code ⸺ by its provisions regarding Crown appeals from jury verdicts ⸺ specifically precludes. It will invite confirmation bias: the verdicts are inconsistent and the jury was misdirected, and so therefore the misdirection must explain away the inconsistency. Just as importantly, it does not discourage the Crown from over‑charging or drafting confusing indictments; if anything, it does the opposite, by eliminating any consequences. None of this is desirable ⸺ and yet, by ignoring the clear requirements of this Court’s judgments and the Criminal Code , this is precisely what my colleagues’ reasons will achieve.”