SUPREME COURT OF CANADA

Citation: R. v. Griffin, 2009 SCC 28, [2009] 2 S.C.R. 42

Date: 20090618 Docket: 32649, 32650

Between:

Her Majesty The Queen Appellant

and

John Griffin Respondent

and between:

Her Majesty The Queen Appellant

and

Earl Roy Harris Respondent

Coram: Binnie, LeBel, Deschamps, Fish, Abella, Charron and Rothstein JJ.

Reasons for Judgment: (paras. 1 to 75)

Joint Dissenting Reasons: (paras. 76 to 111)

Charron J. (Binnie, Deschamps, Abella and Rothstein JJ. concurring) LeBel and Fish JJ.

R. v. Griffin, 2009 SCC 28, [2009] 2 S.C.R. 42

Her Majesty The Queen Appellant

v.

John Griffin Respondent

and

Her Majesty The Queen Appellant

v.

Earl Roy Harris Respondent

Indexed as: R. v. Griffin

Neutral citation: 2009 SCC 28.

File Nos.: 32649, 32650.

2008: November 14; 2009: June 18.

Present: Binnie, LeBel, Deschamps, Fish, Abella, Charron and Rothstein JJ.

on appeal from the court of appeal for quebec

Criminal law — Trial — Charge to jury — Burden of proof — Circumstantial evidence — Reasonable doubt — Whether trial judge erred in his instructions to jury on burden of proof and its application to circumstantial evidence — Whether trial judge erred by suggesting reasonable doubt is doubt for which a reason can be supplied.

Criminal law — Evidence — Hearsay — Admissibility — State of mind exception — Charge to jury — Deceased making statement to his girlfriend shortly before his death indicating that if harm came to him it would be at the hands of her “cousin’s family”, a reference to one of the accused — Whether deceased’s statement admissible under “state of mind” exception to hearsay rule — If so, whether trial judge correctly instructed jury on limited use that could be made of statement.

The accused, G and H, were charged with first degree murder following the shooting death of P. The identity of the killer was the sole issue at trial. P’s girlfriend was the only witness who provided direct evidence on this issue and she identified G as the shooter. P and both accused were all heavily involved in the drug trade. The Crown’s theory was that the shooting was retribution for P’s failure to repay a large drug‑related debt, and that G was the shooter while H acted as the lookout. Circumstantial evidence formed the core of the case against G, and the entirety of that against H. The Crown led evidence that P was driven into hiding and was fearful for his safety in the weeks preceding his death, and that G was on a relentless search for P.

P’s girlfriend testified that shortly before his death, P said to her, “If anything happens to me it’s your cousin’s family.” She understood that P was referring to G, and that he was afraid. Acknowledging the need for an appropriate limiting instruction, the trial judge ruled that the statement was admissible to show the state of mind of P and to rebut the defence proposition that others would have had a motive to kill P at the beginning of 2003.

In his charge to the jury on the burden of proof, the trial judge correctly instructed the jury that, in order to base a verdict of guilt on circumstantial evidence, they had to be convinced beyond a reasonable doubt that the guilt of the accused was the only reasonable inference that could be drawn. However, he also stated that the accused was entitled to an acquittal if the jury found that there was an “equally rational inference” that did not point to guilt, and added that “[i]f there is a second inference that’s as reasonable, you will not be able to base a verdict of guilt on circumstantial evidence.” Following a question from the jury on whether reasonable doubt can be based upon feelings and intuitions, the trial judge repeated his previous instructions, including the ambiguous “equally rational inference” language. He added: “Whether you find there is a reasonable doubt, or whether you find there is not a reasonable doubt, you should individually be in a position to be able to explain your position.” Defence counsel objected to the use of the “equally rational inference” terminology, arguing that it could suggest to the jury that a burden of proof rested on the accused. The trial judge recalled the jury and gave a clarification using the correct language. In providing a limiting instruction to the jury in respect of P’s statement of fear of G, the trial judge told the jury the remark could be used to determine the state of mind of P shortly before his death, and to eliminate other potential people who would want to do P harm, as far as P was concerned.

The jury found G guilty as charged and H guilty of the included offence of manslaughter. The majority of the Court of Appeal set aside the convictions and ordered a new trial on the basis of the trial judge’s errors in his instructions on the burden of proof, and in the limiting instruction given in respect of P’s statement of fear of G.

Held (LeBel and Fish JJ. dissenting): The appeals should be allowed and the convictions restored.

Per Binnie, Deschamps, Abella, Charron and Rothstein JJ.: The trial judge did not err in his instructions to the jury on the burden of proof. The trial judge fulfilled the essential requirement of an instruction on circumstantial evidence by instilling in the jury that in order to convict, they must be satisfied beyond a reasonable doubt that the only rational inference that can be drawn from the circumstantial evidence is that the accused is guilty. Read in context, the words “equally rational” and “as reasonable” were not used as measures of the comparative value or weight of a non‑guilty inference, but to describe the quality of any inference, guilty or otherwise, which might be drawn from the circumstantial evidence. When the instructions on circumstantial evidence are read as a whole, it is more likely than not that they would have been understood by the jury in the manner intended by the trial judge. Moreover, the clarifying recharge was entirely correct. It was sufficient that the trial judge called the jury’s attention to his last instruction on circumstantial evidence, and then corrected the potentially ambiguous “equally rational inference” language with a clear and complete direction. [1] [33‑36] [38]

There is no dispute regarding the correctness of the general instruction on reasonable doubt given in the main charge and repeated in the judge’s reply to the jury’s question. It entirely accords with the Lifchus standard. There is also no error in the additional impugned instruction that an individual juror should be able to explain his or her point of view and identify specific reasons for reasonable doubt. Having regard to the jury charge as a whole and in light of the precise question from the jury on the role of feelings or intuitions in a reasonable doubt analysis, the instruction was simply a caution to jurors to base their verdict not on feelings or intuitions, but on a serious consideration of the evidence. [41‑42] [44]

There is no reason to disturb the trial judge’s ruling on the admissibility of the deceased’s statement made to his girlfriend shortly before his death. That statement was tendered and admitted for the truth of the fact that P himself feared G, a purpose that does not exceed the scope of the traditional “state of mind” exception to the hearsay rule, under which declarations of present state of mind are admissible where the declarant’s state of mind is relevant and the statement is made in a natural manner and not under circumstances of suspicion. In the present case, there was no argument that the statement was made under circumstances of suspicion. P’s fear of G was a relevant fact because P’s fearful state of mind was probative of the nature of the relationship between he and G in the time period preceding the murder. Such information may afford evidence of the accused’s animus or intention to act against the victim, making it relevant to motive and, in turn, to the issue of identification. Accordingly, it was important for the jury to know that P went into hiding and feared for his safety in the weeks preceding his death because of his fear of G, and not because of his fear of someone else, as the defence suggested. The trial judge found that the statement’s probative value outweighed its possible prejudicial effect and admitted the statement into evidence. His decision accorded with the applicable legal principles, and as such, it is entitled to deference. [59] [62‑66] [74]

The trial judge properly instructed the jury not to use P’s statement for the prohibited purpose of proving G’s state of mind, or to conclude that G in fact intended to harm P. He correctly explained that the sole permissible use the jury could make of the statement was as proof of P’s state of mind shortly before his death. He also clearly qualified his assertion that the statement could be used to eliminate other potential people who would want to do P harm with the phrase “as far as P was concerned”. This instruction accurately set out the very purpose for which the statement was tendered: to demonstrate that as far as P was concerned — in his state of mind — G was the only person with cause to do him harm. [71]

Per LeBel and Fish JJ. (dissenting): A new trial should be ordered on the basis of errors in the jury instructions on the burden of proof and of the trial judge’s admission of P’s out‑of‑court statement. The first deficiency in the charge on the burden of proof involved the explanation of how the standard of reasonable doubt applies to circumstantial evidence. Having chosen to elaborate on the basic principle that the jury could return a guilty verdict based on the circumstantial evidence if guilt was the only rational inference they could draw from the evidence, the trial judge should have said that if there was any other (as opposed to an “equally”) rational inference arising from the proven facts that did not point to guilt, the jury would be bound to acquit. Furthermore, in his response to the specific question from the jury as to whether feelings or intuitions could form the basis for reasonable doubt, the trial judge repeated three more times the incorrect component of this instruction, using the term “equally”. Because the jury was likely to have paid particular attention to the answer to their question, this would have amplified the misdirection. In order to ensure that a misdirection to the jury has been successfully rectified, a trial judge will generally have to repeat the direction he or she has given, acknowledge that it was incorrect, tell the jury to put it out of their minds and then tell them in plain and simple terms what the law is. The second deficiency arose in the trial judge’s instruction that the jurors should be able to explain in some detail the basis for any reasonable doubt they might have. While instructing a jury that they should be able to explain their doubts would not necessarily constitute reversible error, here the trial judge’s instructions went one step further, requiring the jury to be able to identify a specific reason for their doubt. In light of either of the uncorrected deficiencies in the charge on the burden of proof, and considering their combined effect in particular, it cannot be said that there is no reasonable possibility the jury was misled in this case. The error was serious. It may well have had significant consequences. The curative proviso under s. 686(1)(b)(iii) of the Criminal Code can therefore not be applied. [77‑80] [83‑87]

The trial judge also erred in admitting the victim’s impugned out‑of‑court statement. The statement was inadmissible hearsay, under either the traditional exceptions or the principled approach. It was not admissible to prove G’s state of mind and support an inference that he had a motive to harm P. The basis for P’s statement that G wished to harm him, and whether P’s perception was accurate, are unknown and purely a matter of speculation. In addition, the statement in question is a statement of opinion which cannot be overcome by any hearsay exception, and there is a substantial risk that it was accorded undue weight: it lacked probative value, and it was highly prejudicial. Similarly, the statement is not admissible to establish P’s state of mind that he was afraid of G. The state of mind of a victim is irrelevant unless it tends to support a permissible inference regarding a relevant fact. Since P’s basis for his belief is not known, a court cannot rely upon this statement. P’s state of mind is therefore irrelevant on its own, and it is impermissible to use it to infer G’s motive. P’s statement was also not admissible to eliminate other suspects, as it could not possibly support an inference that unidentified third parties did not wish him harm. Finally, the statement could not be admitted to support other circumstantial evidence, because its probative value for this purpose is greatly outweighed by its prejudicial effect. This great risk of prejudice could not be attenuated by a limiting instruction. It is difficult to justify admitting the statement for a marginally probative and tangential purpose while insisting that the jury not use it in the most obvious and prejudicial way possible. The decision to admit the statement cannot be characterized as “harmless”, nor can the evidence against the accused be described as “overwhelming”. Accordingly, the curative proviso cannot remedy this error. [77] [88] [97‑99] [102-110]

Cases Cited

By Charron J.

Referred to: R. v. Cooper, 1977 CanLII 11 (SCC), [1978] 1 S.C.R. 860; R. v. Fleet (1997), 1997 CanLII 867 (ON CA), 120 C.C.C. (3d) 457; R. v. Guiboche, 2004 MBCA 16, 183 C.C.C. (3d) 361; R. v. Tombran (2000), 2000 CanLII 2688 (ON CA), 142 C.C.C. (3d) 380; R. v. Lifchus, 1997 CanLII 319 (SCC), [1997] 3 S.C.R. 320; R. v. Naglik, 1993 CanLII 64 (SCC), [1993] 3 S.C.R. 122; R. v. S. (W.D.), 1994 CanLII 76 (SCC), [1994] 3 S.C.R. 521; R. v. Brydon, 1995 CanLII 48 (SCC), [1995] 4 S.C.R. 253; R. v. Seymour, 1996 CanLII 201 (SCC), [1996] 2 S.C.R. 252; R. v. Smith, 1992 CanLII 79 (SCC), [1992] 2 S.C.R. 915; R. v. Starr, 2000 SCC 40, [2000] 2 S.C.R. 144; R. v. P. (R.) (1990), 58 C.C.C. (3d) 334; Lewis v. The Queen, 1979 CanLII 19 (SCC), [1979] 2 S.C.R. 821; R. v. Assoun, 2006 NSCA 47, 244 N.S.R. (2d) 96; R. v. Foreman (2002), 2002 CanLII 6305 (ON CA), 6 C.R. (6th) 201; R. v. Pasqualino, 2008 ONCA 554, 233 C.C.C. (3d) 319; R. v. Lemky (1992), 1992 CanLII 431 (BC CA), 17 B.C.A.C. 71, aff’d 1996 CanLII 235 (SCC), [1996] 1 S.C.R. 757; R. v. Blackman, 2008 SCC 37, [2008] 2 S.C.R. 298; R. v. Couture, 2007 SCC 28, [2007] 2 S.C.R. 517; R. v. Corbett, 1988 CanLII 80 (SCC), [1988] 1 S.C.R. 670.

By LeBel and Fish JJ. (dissenting)

R. v. Cooper, 1977 CanLII 11 (SCC), [1978] 1 S.C.R. 860; R. v. Brydon, 1995 CanLII 48 (SCC), [1995] 4 S.C.R. 253; R. v. Moon, [1969] 3 All E.R. 803; R. v. Seguin (1979), 1979 CanLII 2948 (ON CA), 45 C.C.C. (2d) 498; R. v. Boyd (1999), 1999 CanLII 2870 (ON CA), 118 O.A.C. 85; R. v. Carpenter (1993), 1993 CanLII 8591 (ON CA), 14 O.R. (3d) 641; R. v. Cavanagh (1976), 1976 CanLII 649 (ON CA), 15 O.R. (2d) 173; Shapiro v. Wilkinson, 1943 CanLII 14 (ON CA), [1943] O.R. 806; R. v. Lifchus, 1997 CanLII 319 (SCC), [1997] 3 S.C.R. 320; R. v. Russell, 2000 SCC 55, [2000] 2 S.C.R. 731; R. v. Beauchamp, 2000 SCC 54, [2000] 2 S.C.R. 720; R. v. Gagnon, 2006 SCC 17, [2006] 1 S.C.R. 621; R. v. Blackman, 2008 SCC 37, [2008] 2 S.C.R. 298; R. v. Arp, 1998 CanLII 769 (SCC), [1998] 3 S.C.R. 339; R. v. Mapara, 2005 SCC 23, [2005] 1 S.C.R. 358; R. v. Khan, 1990 CanLII 77 (SCC), [1990] 2 S.C.R. 531; R. v. Starr, 2000 SCC 40, [2000] 2 S.C.R. 144; R. v. Hawkins, 1996 CanLII 154 (SCC), [1996] 3 S.C.R. 1043; R. v. Smith, 1992 CanLII 79 (SCC), [1992] 2 S.C.R. 915; R. v. Khelawon, 2006 SCC 57, [2006] 2 S.C.R. 787; R. v. D.D., 2000 SCC 43, [2000] 2 S.C.R. 275; R. v. Cowell, [2002] O.J. No. 4783 (QL); R. v. Corbett, 1988 CanLII 80 (SCC), [1988] 1 S.C.R. 670; R. v. B. (F.F.), 1993 CanLII 167 (SCC), [1993] 1 S.C.R. 697; R. v. W. (D.), 1991 CanLII 93 (SCC), [1991] 1 S.C.R. 742; R. v. Spence, 2005 SCC 71, [2005] 3 S.C.R. 458; Shepard v. United States, 290 U.S. 96 (1933); R. v. Khan, 2001 SCC 86, [2001] 3 S.C.R. 823.

Statutes and Regulations Cited

Criminal Code, R.S.C. 1985, c. C‑46, s. 686(1)(b)(iii).

Authors Cited

Canada. Law Reform Commission. Report on Evidence. Ottawa: The Commission, 1975.

Paciocco, David M., and Lee Stuesser. The Law of Evidence, 5th ed. Toronto: Irwin Law, 2008.

Sopinka, John, Sidney N. Lederman and Alan W. Bryant. The Law of Evidence in Canada, 2nd ed. Markham, Ont.: Butterworths, 1999.

APPEALS from judgments of the Quebec Court of Appeal (Baudouin, Doyon and Côté JJ.A.), 2008 QCCA 824, 2008 QCCA 825, 237 C.C.C. (3d) 374, 58 C.R. (6th) 86, [2008] Q.J. No. 3589 (QL), [2008] Q.J. No. 3590 (QL), 2008 CarswellQue 3430, setting aside the convictions of both accused and ordering a new trial. Appeals allowed, LeBel and Fish JJ. dissenting.

Thierry Nadon and Geneviève Dagenais, for the appellant.

Louis Belleau, for the respondent Griffin.

No one appeared for the respondent Harris.

The judgment of Binnie, Deschamps, Abella, Charron and Rothstein JJ. was delivered by

Charron J. —

1. Introduction

[1] Following their joint trial for first degree murder before a judge and jury, John Griffin was found guilty as charged and Earl Harris was found guilty of the included offence of manslaughter. Each appealed his conviction on similar grounds, only two of which remain at issue. Both grounds concern the trial judge’s instructions to the jury. The first ground of appeal relates to the instructions on the burden of proof, and the second to the directions on the permissible use of a statement made by the deceased shortly before his death. Doyon J.A. of the Court of Appeal of Quebec (Baudouin J.A. concurring) gave effect to both grounds and ordered a new trial; Côté J.A., in dissent, would have dismissed the appeals: 2008 QCCA 824, 2008 QCCA 825, 237 C.C.C. (3d) 374. The Crown appeals to this Court as of right on both grounds.

[2] For the reasons that follow, I would allow the appeals and restore the convictions.

2. The Evidence at Trial

2.1 The Shooting

[3] Griffin and Harris were charged with first degree murder following the shooting death of Denis Poirier on February 2, 2003. The murder took place at approximately 4:45 p.m., near the intersection of Atwater Avenue and Workman Street in Montreal. Just prior to his death, Poirier and his girlfriend, Jennifer Williams, were eating take-out food in her parked car. Poirier exited the car to urinate in a parking lot on Workman Street. Moments later, he was shot.

[4] The identity of Poirier’s killer was the sole issue at trial. Williams was the only witness who provided direct evidence of the shooter’s identity. She testified that she saw him round the corner of Atwater Avenue and Workman Street, approach Poirier, and fire a weapon three times in his direction, twice after Poirier had gone down. Williams was able to identify the shooter as Griffin.

[5] A second eyewitness to the shooting, Father Yvon Lavoie, observed part of the scene from the window of the presbytery overlooking the parking lot on Workman Street. Father Lavoie saw the profile of a man with a weapon in his right hand. He saw the man fire three shots in a downward direction, but could not see at whom or what the man was shooting. Father Lavoie was unable to identify the shooter from police photos, nor could he identify Griffin as the assailant in court.

[6] The victim and the respondents, Griffin and Harris, were all heavily involved in the drug trade. The Crown’s theory was that the shooting was retribution for Poirier’s failure to repay a large drug-related debt, and that Griffin was the shooter while Harris acted as the lookout. In support of its theory, the Crown led evidence that in the weeks leading up to the killing, Poirier was afraid for his safety and had gone into hiding. During this time, Griffin, at times accompanied by Harris, searched for Poirier with increasing urgency. Apart from the direct evidence of Williams and Father Lavoie, the Crown’s case against Griffin was based on circumstantial evidence. The case against Harris was entirely circumstantial. The circumstantial evidence is summarized in some detail by Doyon J.A. in the Court of Appeal below. The main features of the case against Griffin and Harris are the following.

2.2 The Drug Debt

[7] Matthew Demaine and Joe Besso were drug users and friends of Poirier. Demaine worked at the Claremont, a Montreal restaurant frequented by both Griffin and Poirier.

[8] Demaine testified that on December 27, 2002, he saw Griffin, Harris and a man identified as “Peter the Italian” at the restaurant. He observed Poirier hand over a bag to Peter, during the course of what appeared to be a drug transaction. Sometime after this encounter, Poirier stopped coming to the restaurant. In fact, he left his home and went into hiding at the Clarion Hotel.

[9] In early January 2003, Griffin went to the restaurant. He appeared impatient and told Demaine that he had set up a meeting with Poirier for that day. Poirier never showed up, but later called Demaine, asking him to tell Griffin that everything was all right. Demaine relayed the message.

[10] On January 5, 2003, Peter told Demaine that Poirier owed him an outstanding debt that was to be paid no later than the next day or “he’s done”. Demaine also overheard Peter say to Besso:

This guy is a friend of yours. . . . Well, you tell him he’s got until 5 o’clock tomorrow and if he doesn’t pay me, he can keep the money and he can keep the rest of the shit and I’m going to put a hit on him.

[11] That evening, Demaine and Poirier met in the bar at the Clarion Hotel where Poirier was hiding. Poirier described his financial problems to Demaine. He owed Griffin and Peter approximately $100,000 following drug-related transactions that had gone sour. He explained to Demaine that he could not repay the debt, in part because an intermediary had refused to pay tens of thousands of dollars due to him on the pretext that the quantity and quality of the drug he had provided was unsatisfactory, and because Poirier owed him money.

2.3 Griffin’s Search for Poirier

[12] Following Poirier’s failure to attend the scheduled meeting at the restaurant, Griffin set out on a relentless search for Poirier. He visited the restaurant frequently and called Demaine daily, demanding to know where Poirier was hiding.

[13] Around January 19, 2003, Griffin’s efforts to track down Poirier intensified. He went to the restaurant and gave Demaine “one last chance” to disclose Poirier’s whereabouts. Demaine refused. He was afraid that if Griffin found him, Poirier would be killed.

[14] On January 21, 2003, Poirier telephoned Demaine several times, anxiously asking if he knew where Besso, who had disappeared, could be found. Demaine did not know.

[15] On January 22, 2003, a man arrived at the restaurant and told Demaine, “Johnny [Griffin] wants to talk to you.” Griffin was waiting outside. He ordered Demaine into a vehicle, struck him, and refused to tell him where they were going as the car drove away from the restaurant. The vehicle stopped at a building and Demaine was told to go into the basement. Besso was there and it was obvious that he had been badly beaten. Griffin told Demaine “don’t worry”, he would not be beaten like Besso. Instead, Griffin said that unless he disclosed who had been “dealing” for Poirier and where Poirier was hiding, “I’m going to fucking kill you.” Demaine divulged that Poirier was at the Clarion Hotel. Griffin made a call on his cell phone, telling the person on the other end, “[Y]ou want to do some work tonight? Heavy, pretty heavy”, and then departed, leaving Demaine and Besso sequestered for several hours.

[16] Griffin returned, accompanied by Harris, and ordered Demaine into the same vehicle that had brought him earlier. A man named Ronald Turner drove. The group set out to find Poirier. En route, Griffin ordered Demaine to telephone Poirier and set up a meeting. After several attempts, Demaine was successful in reaching Poirier, but Poirier refused to meet with him. Griffin was furious. He struck Demaine several times, demanding that he persuade Poirier to meet with him by telling him that he feared for his life.

[17] The four men drove first to Poirier’s residence. Griffin and Harris got out and looked for Poirier but he was not there. They next went to the Clarion Hotel. Griffin and Harris entered the hotel, leaving Demaine in the vehicle with Turner. Demaine managed to escape and found refuge in the nearby Ambassador Hotel, where he had the concierge call the police, to whom he relayed the story of his abduction.

[18] That same evening, January 22, 2003, Poirier met with his girlfriend Williams at the bar of the Clarion Hotel. Poirier left the bar at regular intervals throughout the evening to make phone calls and check his messages, and he grew increasingly stressed as the evening went on. Around 11 p.m., Poirier suggested that Williams depart, and as she was leaving he said to her: “If anything happens to me it’s your cousin’s family.” A first cousin of Williams had a child with a first cousin of Griffin, and Williams immediately understood that Poirier was referring to Griffin. Poirier did not explain to Williams why he feared for his safety but she understood he was afraid of Griffin.

[19] The following day, during phone conversations with both Williams and Demaine, Poirier relayed that he had seen Griffin and Harris enter the Clarion Hotel the previous day and had fled to hide in another room. Williams testified that he told her he was “scared for his life” and “didn’t know what to do”.

2.4 The Fatal February 2nd Meeting

[20] In the week preceding the February 2, 2003 shooting, Griffin contacted Poirier’s friend Steven Lauti in an effort to gather information about Poirier’s whereabouts. Lauti did not have any information.

[21] Griffin also met with Jesse McBrearty, who regularly provided Poirier with cocaine, and asked that McBrearty notify him if he planned on meeting with Poirier. McBrearty did not know the nature of the relationship between the two men, only that he had been asked to facilitate a meeting between them.

[22] Shortly thereafter, a meeting was organized between McBrearty and Poirier for February 2, 2003. The two were scheduled to convene at the corner of Atwater Avenue and Workman Street, where Poirier intended to purchase cocaine from McBrearty. McBrearty contacted Griffin and told him of the meeting.

[23] Poirier arrived at the intersection of Atwater Avenue and Workman Street at the designated time on February 2. He was accompanied by Williams, and the two arrived in her car. As McBrearty had not yet appeared, Poirier and Williams went to purchase some take-out food, and then returned to Atwater and Workman. Poirier telephoned McBrearty to see when he would arrive, who in turn called Griffin to relay that Poirier was getting impatient. Griffin told McBrearty: “Don’t worry I’ll be there soon.”

[24] Cellular telephone records indicate that Griffin contacted Harris three times in the moments immediately preceding the murder, and that both men were within a few blocks of the Atwater and Workman intersection at this time. In the moments directly following Poirier’s murder, Griffin again telephoned Harris from the same vicinity. Telephone records further confirm that McBrearty was no where near the site of the shooting at the relevant time.

[25] Upon his arrest on February 5, 2003, Griffin was in possession of the licence plate number of Williams’s car. She had only owned the vehicle for a short time.

[26] Neither Griffin nor Harris testified at trial. The evidence regarding Poirier’s debt to Griffin and the latter’s desire to have the debt repaid was uncontested. The defence argued the existence of a reasonable doubt, alleging that individuals other than Griffin could have wanted to harm Poirier as a result of certain “unfortunate” drug-related transactions in which he had been involved.

3. Analysis

3.1 Instructions to the Jury on the Burden of Proof

[27] The first ground of appeal concerns the instructions to the jury on the burden of proof. The majority in the court below was of the view that the trial judge erred in his instructions on the burden of proof and its application to circumstantial evidence. Although the trial judge correctly instructed the jury on several occasions that, in order to base a verdict of guilt on circumstantial evidence, they had to be convinced beyond a reasonable doubt that the guilt of the accused was the only reasonable inference that could be drawn, the trial judge elsewhere used language suggesting that an inference other than guilt would have to be “as reasonable” in order to lead to an acquittal. In the majority’s view, this language constituted legal error and the corrected instruction given to the jury during the course of their deliberations was insufficient to remedy any potential confusion. The majority further found that the error was compounded by “an excessively demanding instruction to motivate reasonable doubt” (para. 75) given in response to a question from the jury asking whether a reasonable doubt could be based on “feelings, intuitions, et cetera”. Côté J.A. disagreed on both points, finding no legal error.

[28] After explaining the difference between direct and circumstantial evidence and noting that circumstantial evidence formed the core of the case against Griffin, and the entirety of that against Harris, the trial judge gave the following impugned instruction on the burden of proof and circumstantial evidence:

And the point that has to be kept into mind and that is extremely important, because I’ve talked about drawing inferences or drawing rational inferences, and this is something you should note. Before basing a verdict of guilt on circumstantial evidence, so before basing a verdict of guilt on circumstantial evidence, you must be convinced beyond a reasonable doubt that the guilt of the accused is the only reasonable inference to draw from the facts you have chosen to rely upon. I’ll repeat that.

Before basing a verdict of guilt on circumstantial evidence, you must be convinced, beyond a reasonable doubt, that the guilt of the accused is the only reasonable inference to draw from the facts which you have chosen to rely upon.

In other words, if there is an equally reasonable inference, which would indicate an innocent purpose, you could not render a verdict of guilt, because you would not [be] in a position to say, the only reasonable inference that I draw from this circumstantial evidence is guilt. If there is a second inference that’s as reasonable, you will not be able to base a verdict of guilt on circumstantial evidence.

However, after your careful examination of all the evidence, if you arrive at the conclusion, the particular case, the only rationale inference to draw from the circumstantial evidence, which you rely upon, is guilt, then it would be your duty to render a verdict of guilt in those circumstances. [Emphasis added; A.R., at pp. 79-80.]

[29] Counsel made no objection on this point at the conclusion of the main charge.

[30] After two days of deliberation, the jury requested further guidance from the trial judge in this note:

Can we please get a legal definition of reasonable doubt? What things can it be based upon, for example, feelings, intuitions, et cetera? If mostly based on circumstantial evidence, where do we draw the line? Please advise. [A.R., at p. 209]

[31] With the consent of counsel, the trial judge responded to the jury’s question by repeating both the instruction provided in his main charge on the definition of reasonable doubt, which is not in issue here, and the instruction on the application of the burden of proof to circumstantial evidence set out above, this time characterizing an inference other than guilt as an “equally rational inference”. In addition, the trial judge gave a specific response to the suggestion in the jury’s question that a reasonable doubt might be based on feelings or intuitions. I return to the latter component of the judge’s answer to the jury’s question later in these reasons. I deal first with the alleged error in the instructions on circumstantial evidence.

[32] Following the judge’s response to the jury’s question, counsel for the respondents objected to the use of the “equally rational inference” terminology, arguing that it could suggest to the jury that a burden of proof rested on the accused. In the absence of the jury, the trial judge explained to counsel that his use of the phrase “equally rational inference” in this context was equivalent to saying “another reasonable inference”. Nevertheless, out of an abundance of caution, he recalled the jury and gave the following clarification:

Just on the last point when dealing with circumstantial evidence and reasonable doubt and the way the two (2) meld together, I just want to make sure that it’s clear in your mind that in order to render a verdict of guilt, based upon circumstantial evidence, the guilt of the accused would have to be the only rational inference that you could draw from the circumstantial evidence.

If there’s any other rational inference that you can draw from the evidence that would mean that you would not be able to render a verdict of guilt based on the circumstantial evidence, because then there would be — at best you would have reached maybe a probability or a likelihood of guilt, but not have reached that higher standard of proof beyond a reasonable doubt. [A.R., at p. 216]

[33] We have long departed from any legal requirement for a “special instruction” on circumstantial evidence, even where the issue is one of identification: R. v. Cooper, 1977 CanLII 11 (SCC), [1978] 1 S.C.R. 860. The essential component of an instruction on circumstantial evidence is to instill in the jury that in order to convict, they must be satisfied beyond a reasonable doubt that the only rational inference that can be drawn from the circumstantial evidence is that the accused is guilty. Imparting the necessary message to the jury may be achieved in different ways: R. v. Fleet (1997), 1997 CanLII 867 (ON CA), 120 C.C.C. (3d) 457 (Ont. C.A.), at para. 20. See also R. v. Guiboche, 2004 MBCA 16, 183 C.C.C. (3d) 361, at paras. 108-10; R. v. Tombran (2000), 2000 CanLII 2688 (ON CA), 142 C.C.C. (3d) 380 (Ont. C.A.), at para. 29.

[34] There is no question that the instructions in the present case fulfilled this essential requirement. The trial judge repeatedly made clear to the jury that a guilty verdict can only be rendered if guilt is the sole rational inference to be drawn from the circumstantial evidence. The issue to be resolved, rather, is whether the use of the words “equally rational” and “as reasonable” detracted from the otherwise correct instructions by leading jurors to believe, as the respondents contend, that they could only acquit if they found an innocent inference from the circumstantial evidence of the same weight as an inference of guilt. It is argued that the impugned language had the potential to engage the jury in an abstract comparative exercise, assessing the qualitative reasonableness of one inference against another when the mere existence of any rational, non-guilty inference is sufficient to raise a reasonable doubt.

[35] Read in context, as all jury instructions must be, it becomes apparent that the words “equally rational” and “as reasonable” were not used as measures of the comparative value or weight of a non-guilty inference, but, as the judge himself explained to counsel, to describe the quality of any inference, guilty or otherwise, which might be drawn from the circumstantial evidence. Indeed, no one argues that the jury should be entitled to base its decision on irrational or unreasonable inferences. When the instructions on circumstantial evidence are read as a whole, it is more likely than not that they would have been understood by the jury in the manner intended by the trial judge. That said, it is my view that the instructions on this point were not as clear as they could have been and, consequently, the trial judge was wise to recharge the jury as he did.

[36] Both parties and the court below, unanimous on this point, agreed that the clarifying recharge was entirely correct. Griffin and Harris argue, however, that this final charge could not have rectified any possible confusion that may have resulted from the use of the impugned language because the trial judge failed to refer back to the problematic phrases employed in the earlier instructions. This argument was accepted by the majority of the Court of Appeal, and rejected by Côté J.A.

[37] Griffin suggests that the trial judge should have instructed the jury in terms akin to the following:

What was required to correct the deleterious effect of the defective instructions was for the trial judge to tell the jury that they didn’t need to determine whether any innocent inference they may consider was more reasonable or as reasonable, or less reasonable than an inference of guilt and that, as long as an inference not indicating guilt could be considered rational or reasonable and was not inconsistent with the evidence, they must acquit. [R.F., at para. 52]

[38] With respect, I agree with Côté J.A. that an instruction along the lines proposed by Griffin would serve only to exacerbate any potential confusion stemming from the previous instructions. It was sufficient that the trial judge called the jury’s attention to his last instruction on circumstantial evidence, and then corrected the potentially ambiguous “equally rational inference” language with a clear and complete direction.

[39] I now turn to the trial judge’s answer to the jury’s question whether a reasonable doubt could be based on feelings or intuitions. After canvassing the gist of his proposed answer with counsel and obtaining their approval, the trial judge instructed the jury as follows:

So, when you ask in your question, is it feelings, is it intuitions, I would answer no to that question. Whether you find there is a reasonable doubt, or whether you find there is not a reasonable doubt, you should individually be in a position to be able to explain your position. You should be able to say, for instance, “Listen, I have a reasonable doubt on this essential element, and here’s why I have a reasonable doubt. I can’t rely on Witness A and B. I don’t find Witness A and B credible. And here’s why I don’t find Witness A and B credible, because there’s contradictions, or they’re not confirmed in the evidence, and I don’t accept that Exhibits X, Y and Z help me out on this particular issue,” so that you’re able as a judge to explain why you are supporting a particular position.

On the other hand, you should be able, also, if your position is, “I don’t have a reasonable doubt on this issue. I think the Crown has proved this essential element” to be able to explain it, to be able to say, “The reason I’ve arrived at that decision,” for example, “is that I accept the testimony of Witnesses A and B. I saw them testify. I saw the manner in which they testified. I find that there is confirmation of their testimony within the evidence. I also rely upon Exhibits X, Y and Z, because it adds -- the facts that I garnered out of Exhibits X, Y and Z, reinforce my opinion that the Crown has proved that particular essential element beyond a reasonable doubt.”

So, as you can see, it’s really a cold objective analysis of the facts, but it’s based upon the facts, and you might arrive at the conclusion as “I have a reasonable doubt, because there’s a lack of evidence on a particular issue that I think is important in my mind,” or there’s a contradiction in the evidence.

So, it’s not feelings or intuitions, and then -- either way, I mean, you can’t say, “Oh, I just got a feeling that the person is guilty.” That means that you’re not really analyzing the evidence. You should be able to explain beyond simply saying, “Well, I’ve got a feeling, or I’ve got an intuition”, because you’re called upon to base your decision upon the facts that have been -- upon the proof that’s been presented to you. [A.R., at pp. 211-13]

[40] Counsel raised no objection to this instruction.

[41] On appeal in the court below Griffin and Harris took issue with the suggestion in the above instruction that an individual juror should be able to explain his or her point of view and identify specific reasons for his or her reasonable doubt. They argued that this instruction tended to shift the burden of proof. The majority accepted this argument, concluding that in combination with the judge’s failure to adequately correct the ambiguous “equally rational” and “as reasonable” terminology, this “excessively demanding instruction to motivate reasonable doubt, rendered the trial unfair” (para. 75).

[42] Côté J.A. disagreed with the majority and found no error. Having regard to the jury charge as a whole and in light of the precise question from the jury on the role of feelings or intuitions in a reasonable doubt analysis, the impugned instruction was simply a caution to jurors to base their verdict not on feelings or intuitions, but on a serious consideration of the evidence. Côté J.A. stated the following (at para. 139):

[translation] Although it is preferable not to invite jurors to motivate their reasonable doubt, one fact remains, that the standard requires a serious examination of the evidence upon which the verdict is to be based. I do not believe that this instruction led the jurors to apply a standard of proof which is lower than that of proof beyond a reasonable doubt.

[43] As the majority of the Court of Appeal rightly acknowledged (at para. 65), this Court in R. v. Lifchus, 1997 CanLII 319 (SCC), [1997] 3 S.C.R. 320, “did not entirely reject a definition of reasonable doubt that would include a reference to reasons”. In settling the preferred approach to a jury charge on reasonable doubt in Lifchus, Cory J. noted the appellate controversy on whether a jury should be instructed that a reasonable doubt is a doubt “for which one can give a reason” (para. 28). After considering the potential risks and difficulties that accompany such an instruction, he concluded at para. 30:

It follows that it is certainly not essential to instruct jurors that a reasonable doubt is a doubt for which a reason can be supplied. To do so may unnecessarily complicate the task of the jury. It will suffice to instruct the jury that a reasonable doubt is a doubt based on reason and common sense which must be logically based upon the evidence or lack of evidence.

[44] In my view, the trial judge’s instruction in this case accords entirely with the Lifchus standard. There is no dispute regarding the correctness of the general instruction on reasonable doubt given in the main charge and repeated in the judge’s reply to the jury’s question. Among other things, the general reasonable doubt instruction included the explanation suggested in Lifchus that “[a] reasonable doubt is based upon reason and common sense. It must logically arise from the evidence or a contradiction in the evidence or an absence of the evidence which you have heard during this trial.” In further compliance with Lifchus, the trial judge did not “unnecessarily complicate the task of the jury” in the main charge by suggesting that a reasonable doubt is a doubt for which a reason can be supplied. Indeed, at one point in his charge, he stressed that the jury would “not be called upon to justify your final verdict to anyone”, telling the jurors:

You will not have to explain why you arrived at a particular verdict. You will not have to stand up and say guilty because or not guilty because. . . .

Furthermore, no one is allowed to ask you, even myself, what happened during your deliberations, what went on, what was the thought process? [A.R., at p. 65]

The substance of the main charge on reasonable doubt was therefore entirely correct.

[45] The jury’s question was a clear indication that it still required assistance on the definition of reasonable doubt and the application of this concept to the evidence. This Court has repeatedly stressed the importance of providing clear, precise answers to questions from a jury: R. v. Naglik, 1993 CanLII 64 (SCC), [1993] 3 S.C.R. 122, at p. 139; R. v. S. (W.D.), 1994 CanLII 76 (SCC), [1994] 3 S.C.R. 521, at pp. 528-31; R. v. Brydon, 1995 CanLII 48 (SCC), [1995] 4 S.C.R. 253, at paras. 16 and 19; and R. v. Seymour, 1996 CanLII 201 (SCC), [1996] 2 S.C.R. 252, at para. 30. Given the specificity of the query, it was reasonable for the trial judge to decide that more was required than a simple reiteration of the instructions contained in the main charge. In particular, the trial judge recognized that it was important to tell the jury that their verdict must be based, not on feelings or intuitions but rather, as set out in Lifchus above, “on reason and common sense” which, in turn, “must be logically based upon the evidence or lack of evidence” before the court (emphasis added). The trial judge explained what that meant by setting out generic examples of how a juror’s analysis on reasonable doubt might proceed.

[46] The majority in the court below was of the view that “[s]tanding alone” the suggestion that individual jurors be in a position to explain their conclusion on reasonable doubt “would not necessarily have been problematic” (para. 59). The majority took exception with the examples that the trial judge gave in explaining a reasonable doubt analysis because these examples exacted too demanding a standard for the jury. Doyon J.A. explained his view as follows at para. 60:

[E]ach of the jurors having a reasonable doubt would have to be in a position to explain his or her point of view and identify specific reasons for his or her reasonable doubt. This is the meaning of the very demanding examples given by the trial judge, further to which the jurors would have to be able to say, for instance, that they did not believe a witness and, in addition, be able to explain precisely why they did not believe him.

[47] I respectfully disagree. In my view, the instruction contains no error. As Côté J.A. put it, “the standard requires a serious examination of the evidence upon which the verdict is to be based” (para. 139) and that was the essence of the message conveyed to the jury. It is particularly noteworthy that the trial judge in this case wisely sought the advice of counsel before answering the jury’s question. Questions that arise during the course of jury deliberations are often crucially important and are not always easy to answer. Counsel are well placed to provide valuable assistance to a trial judge in crafting appropriate answers to jury questions. Counsel in this case raised no objection, either to the trial judge’s proposed answer, or to the instructions themselves after they were delivered. One can only conclude that they could perceive no difficulty with the trial judge’s answer in the context it was given. In these circumstances, their argument on appeal that the trial judge could have better captured the essence of the definition of reasonable doubt by using different examples resonates hollow.

3.2 Poirier’s Statement to Williams

[48] The sole statement at issue in this appeal is Poirier’s statement to Williams, “[i]f anything happens to me it’s your cousin’s family.” Griffin and Harris argued in the court below that this statement should not have been admitted into evidence by the trial judge. The Court of Appeal was divided not on the question of admissibility, but on the likely effect of the trial judge’s limiting instruction to the jury on the use that could be made of the impugned statement. In this Court, Griffin and Harris argue in support of the majority’s conclusion below that the limiting instruction was insufficient, and also repeat their primary argument that the trial judge erred by admitting Poirier’s statement to Williams in the first place. I will deal first with the question of admissibility before commenting on the limiting instruction.

3.2.1 The Admissibility of Poirier’s Statement to Williams

[49] Assessing the admissibility of Poirier’s statement, “[i]f anything happens to me it’s your cousin’s family”, necessitates a review of what transpired at trial. As is plain from the summary of the evidence provided at the outset of these reasons, both Demaine and Williams testified as to several statements they received from Poirier while he was in hiding in the weeks leading up to his murder. Poirier not being available for cross-examination, these statements of course constitute hearsay if introduced for the proof of their contents. Since several statements the Crown sought to tender at trial arguably fell into the hearsay category, Brunton J. held a voir dire at the commencement of the trial, based on the preliminary hearing transcripts of the witnesses’ testimony, to determine the admissibility of numerous statements made by Poirier to Demaine and Williams.

[50] The Crown argued on the voir dire that several of Poirier’s statements explained his state of mind or conduct at the time they were spoken and were therefore admissible under the traditional “state of mind” or “present intentions” exception to the hearsay rule. The Crown also relied on the principled exception to the hearsay rule based on the twin criteria of necessity and reliability.

[51] Counsel for Griffin and counsel for Harris made no submissions on the voir dire regarding the traditional exception to the hearsay rule, but contested the admission of the statements under the principled exception. Counsel for Griffin argued that Demaine’s credibility was questionable, but quickly conceded that this issue could be fully canvassed at trial and was not of concern at the stage of establishing threshold reliability. He presented no other argument. Counsel for Harris raised concerns about the reliability of Williams’ testimony in respect of Poirier’s statement, “[i]f anything happens to me it’s your cousin’s family” based on an alleged statement wherein Poirier had told another witness that Griffin had indicated he no longer wanted to kill him. Counsel acknowledged, however, that this other statement was not before the court.

[52] The trial judge ruled the statements admissible. The following excerpt of his reasons on the voir dire demonstrates that he was alive to the very concern that now occupies this Court:

While the issue was not raised by the defence, I have directed my attention particularly to the statement received by Ms. Williams from Mr. Poirier to the effect that if anything happened to him, it would be due to her cousin’s family. As I have mentioned, Ms. Williams will provide evidence which links this phrase to Mr. Griffin.

As pointed out in Smith and Starr, the “state of mind” or “present intentions” traditional exception to the hearsay rule does not permit the admissibility of hearsay to show the state of mind of a third party. The same holds true if a statement is admitted pursuant to the principled exception to the hearsay rule.

In Starr, Iacobucci J. would have declared the proposed hearsay of the declarant inadmissible for a number of reasons. One was based upon the fact that the trier of fact was asked to infer Starr’s intentions by use of the declarant’s hearsay statement. Iacobucci J. felt that in such a case, the prejudice occasioned to Starr outweighed the probative value of the evidence.

I have asked myself whether the same situation exists in the present case. Is it not tempting to leap from the fact that Poirier’s state of mind is to the effect that if anything happened to him it would be due implicitly to Griffin to an inference that his death was in fact caused by Griffin?

If the evidence was not relevant to some other issue in the trial, it might well be that the prejudicial nature of the evidence would outweigh its probative value. I am of the opinion that the evidence does address another issue.

During the voir dire, I was provided the testimony of Ms. Williams rendered during the preliminary inquiry. During her cross‑examination, questions were posed in an attempt to show that Mr. Poirier was concerned about his safety well before he had had any dealings with Mr. Griffin. No doubt the questions were posed to establish that someone other than John Griffin had a motive to kill Poirier.

The hearsay statement of Poirier pointing to his belief, shortly before his death, that if harm came to him it would be at the hands of Griffin is relevant to rebut the notion that others would also have had a motive to kill him at the beginning of 2003.

In these circumstances, while appropriate jury instructions will have to be given, I am of the opinion that the probative value of the hearsay evidence of Poirier, which implicitly identifies Mr. Griffin as someone who wishes him harm, outweighs any prejudicial effect.

(2005 CanLII 5629, at paras. 26-33)

[53] In this Court, Griffin and Harris argue that the trial judge ought to have ruled Poirier’s statement to Williams inadmissible because it “contained no expression whatsoever of his own intentions”, nor did it indicate that he feared Griffin. In the respondents’ submission, the statement, “[i]f anything happens to me it’s your cousin’s family” expressed nothing but Poirier’s belief that Griffin intended to harm him. As such, Poirier’s was a statement about Griffin’s intentions, not his own, and it was therefore inadmissible hearsay. Alternatively, Griffin and Harris argue that if the statement was admissible as expressive of Poirier’s state of mind, the prejudicial effect of the statement outweighed its probative value.

[54] I will deal firstly with the argument that the impugned statement was one of Griffin’s “present intentions”, not Poirier’s and, as such, fell outside the scope of the traditional exception to the hearsay rule.

[55] As the trial judge noted in his ruling on the voir dire, the danger inherent in admitting Poirier’s statement into evidence was that the jury might be tempted to leap from the fact of Poirier’s fear of Griffin to the conclusion that Griffin in fact intended to kill him, and ultimately carried out the murder. In other words, the concern was that the jury might take the statement as proof of Griffin’s intentions. Citing both R. v. Smith, 1992 CanLII 79 (SCC), [1992] 2 S.C.R. 915, and R. v. Starr, 2000 SCC 40, [2000] 2 S.C.R. 144, the trial judge correctly observed that the “state of mind” or “present intentions” exception to the hearsay rule does not permit the admission of hearsay to show the state of mind of a third party. This principle requires some further elaboration in the context of this case.

[56] In Smith, Lamer C.J. cited with approval at p. 927, the following excerpt from R. v. P. (R.) (1990), 58 C.C.C. (3d) 334 (Ont. H.C.J.), per Doherty J. (as he then was), at pp. 343-44:

An utterance indicating that a deceased had a certain intention or design will afford evidence that the deceased acted in accordance with that stated intention or plan where it is reasonable to infer that the deceased did so. . . .

The rules of evidence as developed to this point do not exclude evidence of utterances by a deceased which reveal her state of mind, but rather appear to provide specifically for their admission where relevant. The evidence is not, however, admissible to show the state of mind of persons other than the deceased (unless they were aware of the statements), or to show that persons other than the deceased acted in accordance with the deceased’s stated intentions, save perhaps cases where the act was a joint one involving the deceased and another person. [Emphasis added.]

[57] Writing for the majority in Starr, Iacobucci J. referred to the above passage in Smith and elaborated on the reasons why a statement of intention cannot be admitted to prove the intentions of someone other than the declarant, stating in part at para. 172:

[T]here are very good reasons behind the rule against allowing statements of present intention to be used to prove the state of mind of someone other than the declarant. As noted above, the central concern with hearsay is the inability of the trier of fact to test the reliability of the declarant’s assertion. When the statement is tendered to prove the intentions of a third party, this danger is multiplied. If a declarant makes a statement about the intentions of a third party, there are three possible bases for this statement: first, it could be based on a prior conversation with the accused; second, it could be based on a prior conversation with a fourth party, who indicated the third party’s intentions to the declarant; or third, it could be based on pure speculation on the part of the declarant. Under the first scenario, the statement is double hearsay. Since each level of double hearsay must fall within an exception, or be admissible under the principled approach, the mere fact that the declarant is making a statement of present intention is insufficient to render it admissible. The second level of hearsay must also be admissible. [Emphasis in original deleted.]

[58] Applying Smith and Starr to the facts of the present case, there is no doubt that Poirier’s statement cannot be admitted as proof of Griffin’s intentions, because we do not know the basis on which Poirier came to believe that if he was harmed, Griffin would be responsible. Hence, Poirier’s statement is not admissible to prove a third party’s intentions, unless the “second level of hearsay” is also admissible. The same principle applies to statements of present state of mind. If, for example, Poirier had said to Williams that his friend Besso was afraid of Griffin, the statement could not be admitted to prove that Besso in fact feared Griffin. That the deceased’s hearsay statement about his own state of mind or intention to follow a particular course of action cannot constitute proof of another person’s state of mind or intentions only makes good sense.

[59] In the case before us, however, the statement at issue was not admitted to prove the state of mind or intentions of a third party. No one questioned at trial, or in the court below, that the inference that Poirier feared Griffin could be drawn from the former’s statement to Williams. The statement was tendered and admitted for the truth of the fact that Poirier himself feared Griffin, a purpose that does not exceed the scope of the “state of mind” exception to the hearsay rule. As this Court stated in Starr, declarations of present state of mind are admissible under the traditional exception to the hearsay rule where the declarant’s state of mind is relevant and the statement is made in a natural manner and not under circumstances of suspicion (para. 168). In the present case, there was no argument that the statement was made under circumstances of suspicion. Poirier’s fear of Griffin was a relevant fact — it was relevant to motive and, in turn to the issue of identification. Although a declarant’s hearsay statement cannot constitute proof of a third party’s state of mind, this does not mean that the declarant’s state of mind can have no bearing on other issues in the case. I will explain.

[60] As stated earlier, the sole issue at trial was the identity of Poirier’s killer. Although Poirier’s state of mind may have no direct bearing on the resolution of the identity of his murderer, it is well established that a deceased’s mental state may be relevant to the question of an accused’s motive. As Doherty J. explained in the oft‑approved judgment in P. (R.), at p. 339:

. . . the deceased’s mental state may bear no direct relevance to the ultimate issue of identification but it will none the less be relevant to that issue if it is relevant to another fact (e.g., motive) which is directly relevant to the ultimate issue of identification.

In turn, that evidence of motive is relevant and admissible particularly where, on the issues of identity and intention, the evidence is purely circumstantial, is equally well established at law: Lewis v. The Queen, 1979 CanLII 19 (SCC), [1979] 2 S.C.R. 821, at pp. 834-38.

[61] The state of the relationship between a deceased and an accused in the time period leading up to the former’s murder has been recognized as probative of the issue of motive. For example, in R. v. Assoun, 2006 NSCA 47, 244 N.S.R. (2d) 96, identity was the crucial issue at trial. The trial judge admitted statements made by the deceased expressing fear of the accused on the basis that such statements were probative of the issues of “malice, motive, [and] state of mind” (para. 104) which in turn were relevant to identity. The Nova Scotia Court of Appeal agreed (at para. 133):

[The deceased’s] state of mind and [the accused’s] state of mind are probative of the relationship between them at the time of [the deceased’s] murder. Therefore, they are probative of motive, which is relevant to identity.

[62] The conclusion in Assoun echoes that of the Ontario Court of Appeal in R. v. Foreman (2002), 2002 CanLII 6305 (ON CA), 6 C.R. (6th) 201, where the court considered the relationship between a deceased’s state of mind and the issue of motive. Upholding the trial judge’s admission of statements made by the deceased shortly before her death in which she expressed fear of the accused, the court stated as follows (at para. 30):

Motive refers to an accused’s state of mind. . . . [T]he deceased’s state of mind was one link in a chain of reasoning which could lead to a finding that the [accused] had a motive to kill [the deceased]. In that way, evidence of [the deceased’s] state of mind had an indirect connection to the appellant’s state of mind.

[63] The connection between a deceased’s state of mind and that of an accused arises by virtue of a pre-existing relationship between the two; if a deceased and an accused are unknown to one another, this course of logic can find no application. That the relationship between a deceased and an accused was acrimonious or that the two had engaged in a dispute in the period leading up to a murder are highly relevant to the issue of motive because such information may afford evidence of the accused’s animus or intention to act against the victim: R. v. Pasqualino, 2008 ONCA 554, 233 C.C.C. (3d) 319, at para. 31. See also R. v. Lemky (1992), 1992 CanLII 431 (BC CA), 17 B.C.A.C. 71, aff’d 1996 CanLII 235 (SCC), [1996] 1 S.C.R. 757. This is not to say that a deceased’s state of mind alone is capable of proving motive. Insofar as it affords evidence of the nature of the relationship between a deceased and an accused, however, a deceased’s state of mind is one piece of evidence that may be relevant to the issue of motive.

[64] In the instant appeal, Poirier’s fearful state of mind is probative of the nature of the relationship between he and Griffin in the time period immediately preceding the murder. The two were known to one another prior to the murder and had engaged in drug-related transactions resulting in a sizeable debt owed to Griffin by Poirier. Poirier sensed that Griffin’s animus towards him was increasing throughout January 2003 and it is apparent that he acted accordingly: Poirier refused to meet with Griffin, he went into hiding at the Clarion Hotel, and, by way of the statement now at issue before this Court, he told his girlfriend that he believed that Griffin intended him harm. All of these facts make it more likely that Griffin had the motive ascribed to him by the Crown than would otherwise be the case. Indeed, quite apart from this statement, the evidence that Griffin had a motive to harm Poirier was overwhelming.

[65] In addition to its broad relevance as one piece of a large puzzle of circumstantial evidence going to proof of motive, Poirier’s state of mind took on an elevated importance in light of the defence argument advanced in this case. In admitting the impugned statement, the trial judge correctly observed that Poirier’s particularized fear of Griffin was evidence that tended to rebut the defence proposition that someone other than Griffin might have had a motive to kill Poirier in January 2003. While Poirier’s fear of Griffin was not conclusive on this issue, the evidence was highly relevant to the question of identity. That Poirier was driven into hiding and was fearful for his safety in the weeks preceding his death because of his fear of Griffin, and not because of his fear of someone else as suggested by the defence, could be properly considered as one piece of circumstantial evidence refuting the defence theory that someone other than Griffin may have killed Poirier.

[66] Thus, the trial judge correctly appreciated that the deceased’s mental state was highly probative of the issue of motive and that it took on additional probative value in that it tended to rebut the specific defence proposition that someone other than Griffin may have had a motive to kill Poirier in January 2003. Properly acknowledging the need for an appropriate limiting instruction, the trial judge found that the statement’s probative value outweighed its possible prejudicial effect and admitted the statement into evidence. The trial judge’s decision accorded with the applicable legal principles, and as such, it is entitled to deference: R. v. Blackman, 2008 SCC 37, [2008] 2 S.C.R. 298, at para. 36, and R. v. Couture, 2007 SCC 28, [2007] 2 S.C.R. 517, at para. 81. I see no reason to disturb the trial judge’s ruling on admissibility.

3.2.2 The Limiting Instruction

[67] I now turn to the limiting instruction given to the jury in respect of Poirier’s statement of fear of Griffin. The trial judge first told the jury that it should approach the evidence regarding statements by Poirier to others with caution given that Poirier was not under oath, nor present in court to be observed or cross-examined. An instruction to the same effect was also given during the course of the trial at the time the various hearsay statements were introduced in evidence. The trial judge informed the jurors that they should carefully assess the credibility of those witnesses recounting utterances by Poirier before deciding to place any reliance on such statements. The trial judge then gave the following directions on the permissible uses to which Poirier’s statement to Williams could be put:

That remark can be used, one, to determine the state of mind of Mr. Poirier shortly before his death, that he felt, his state of mind was, that if he was in any danger, it would be due to Mr. Griffin. It cannot be used as proof of the state of mind of Mr. Griffin. It cannot be used by you to arrive at the conclusion that indeed Mr. Griffin had a state of mind where he wanted to cause harm to Mr. Poirier.

That state of mind of Mr. Poirier, if you accept that evidence, that he felt at that particular moment in time, that if he was in danger, it was because of Mr. Griffin, can address another issue. There was cross‑examination, both of madam Williams and of other witnesses, which show that beginning in the autumn of two thousand and two (2002), Mr. Poirier had some security concerns. You will recall that madam Williams left their common residence to go live on Cavendish Street in the autumn, well before it would appear -- that Mr. Poirier would have had any dealings or that the event that would have lead to the conflict with Mr. Griffin would have occurred. There is also evidence that he would have had business problems with other people, Mr. Jeobbey, the [h]ash deal, where he would have owed money to Mr. Jeobbey.

So, there is proof for you to consider that perhaps Mr. Poirier had other people who weren’t very happy with him. But you can use, if you accept that he said it, when he tells madam Williams, shortly before his death: If anything happens to me, it’s going to be due to your cousin’s family. You can use that to eliminate other potential people who would want to do him harm, as far as he was concerned. In other words, as far as he was concerned, shortly before his death, notwithstanding the problems he had with other people, he felt that the only source of real problems for him was his cousin’s family, but that goes again to his state of mind. And I emphasize again that you cannot use that evidence to impute a state of mind, to give a state of mind to Mr. Griffin. You cannot use that statement to say, Mr. Griffin wanted to cause him harm. [Emphasis added; A.R., at pp. 104-6.]

[68] Doyon J.A. for the majority of the Court of Appeal was of the opinion that this instruction properly limited the statement for use in establishing Poirier’s state of mind, but not Griffin’s. Indeed, he stated, “[i]f the trial judge had limited the use of the statement to establish that Mr. Poirier was afraid of Mr. Griffin, I would have rejected this ground of appeal” (para. 90). Doyon J.A.’s difficulty arose from the second part of the instruction, where the trial judge told the jury that it could use the statement to “eliminate other potential people who would want to do him harm, as far as he was concerned”. Doyon J.A. was of the opinion that there existed “a reasonable likelihood that the jury improperly used the statement of the deceased to actually eliminate other potential murderers, instead of limiting this conclusion to the victim’s state of mind” (para. 90). In his view, “[t]he distinction between actually eliminating other potential murderers and limiting this conclusion to the victim’s state of mind is so tenuous that it is virtually impossible to conclude that the jury applied it” (para. 89).

[69] Côté J.A. disagreed, finding that the limiting instruction was correct in law, and that the jury must be trusted to have followed the judge’s directions. As she put it (at para. 174): [translation] “In my view, the intelligence of jurors should be trusted as to their comprehension of instructions concerning the limited use which they can make of this evidence”: R. v. Corbett, 1988 CanLII 80 (SCC), [1988] 1 S.C.R. 670, at p. 695.

[70] I agree with Côté J.A.

[71] The trial judge properly instructed the jury not to use Poirier’s statement for the prohibited purpose of proving Griffin’s state of mind, or to conclude that Griffin in fact intended to harm Poirier. He correctly explained that the sole permissible use the jury could make of the statement was as proof of Poirier’s state of mind shortly before his death. He also clearly qualified his assertion that the statement could be used to “eliminate other potential people who would want to do him harm” with the phrase “as far as he was concerned”, “he” being Poirier. This instruction accurately set out the very purpose for which the statement was tendered — to demonstrate that as far as Poirier was concerned — in his state of mind — Griffin was the only person with cause to do him harm. After making this statement, the trial judge immediately re-emphasized that the jury could not use Poirier’s statement to impute a state of mind to Griffin. The trial judge did not invite the jury to use the statement for a prohibited purpose. Accordingly, the instructions contained no error.

[72] In my respectful view, the majority erred in finding that the distinction drawn by the trial judge in his limiting instruction “between actually eliminating other potential murderers and limiting this conclusion to the victim’s state of mind” was “so tenuous that it is virtually impossible to conclude that the jury applied it” (para. 89). To make too much of the risk that the jury might misuse evidence is contrary to established principles of law regarding jury trials. As Côté J.A. aptly noted, juries must be trusted to have the requisite intelligence to perform their duties in accordance with the instructions given to them by the trial judge. Dickson C.J. made this point quite forcefully in Corbett, at p. 692. His words are entirely apposite here:

In my view, it would be quite wrong to make too much of the risk that the jury might use the evidence for an improper purpose. This line of thinking could seriously undermine the entire jury system. The very strength of the jury is that the ultimate issue of guilt or innocence is determined by a group of ordinary citizens who are not legal specialists and who bring to the legal process a healthy measure of common sense. The jury is, of course, bound to follow the law as it is explained by the trial judge. Jury directions are often long and difficult, but the experience of trial judges is that juries do perform their duty according to the law. We should regard with grave suspicion arguments which assert that depriving the jury of all relevant information is preferable to giving them everything, with a careful explanation as to any limitations on the use to which they may put that information. So long as the jury is given a clear instruction as to how it may and how it may not use evidence of prior convictions put to an accused on cross‑examination, it can be argued that the risk of improper use is outweighed by the much more serious risk of error should the jury be forced to decide the issue in the dark. [Emphasis in original deleted.]

[73] Dickson C.J. went on to note that in many situations — for example, when dealing with similar fact evidence or prior inconsistent statements — the jury is permitted to hear and use evidence relevant to one issue, but not to another. All that is required in each case is a clear direction to the jury indicating the permissible and impermissible uses to which the evidence may and may not be put. Dickson C.J. explained, at p. 695:

If risk that the jury might misuse evidence were enough to render such evidence inadmissible in all cases, then in each of the situations just identified, the evidence would have to be excluded. Yet the risk of error inherent in depriving the jury of such information is so strong that the balance is struck by allowing the evidence to be received, subject to the trial judge’s discretion, but at the same time insisting on a careful direction from the trial judge as to the permissible conclusion or inferences which may be drawn.

[74] In the present case, it was important for the jury to know that Poirier went into hiding and feared for his safety not because of some other drug deal that may have turned sour or because he feared the police, as was suggested in the course of the trial, but because he feared Griffin. To leave the jury in the dark on this important feature of the case would distort the truth seeking function of the trial. Furthermore, there was nothing particularly complex about the limited use the jury could make of this evidence. That a statement made by Poirier, about the person he feared at the time, could only go to establishing his state of mind and not that of Griffin or anyone else, is at its core a proposition that entirely accords with common sense, the very attribute which gives the jury its strength.

4. Disposition

[75] I would allow the appeals and restore the convictions.

The following are the reasons delivered by

LeBel and Fish JJ. (dissenting) —

I. Introduction

[76] The relevant facts are fully and accurately set out in the reasons of Charron J., and we have nothing to add in that regard. With respect, however, we are of a different opinion and have arrived at a different result. We would dismiss the Crown’s appeals and confirm the order of the Court of Appeal for a new trial for both respondents.

[77] We agree with Doyon J.A. (Baudouin J.A. concurring) that a new trial must be ordered on the basis of errors in the jury instructions on the burden of proof. Doyon J.A. also concluded that the trial judge should have limited more strictly the jury’s use of an out-of-court statement made by the victim, Denis Poirier, about his fear of the respondent, John Griffin. In our view, that statement should not have been put before the jury in the first place. It was inadmissible hearsay, under either the traditional exceptions or the principled approach. The admission of the statement likewise requires a new trial.

II. Analysis

A. The Instructions on the Burden of Proof

[78] There were two substantial deficiencies in the trial judge’s charge on the burden of proof. The first involved the explanation of how the standard of reasonable doubt applies to circumstantial evidence. The second arose from the instruction that the jurors should be able to explain in some detail the basis of any reasonable doubt they might have.

[79] In his original charge, the trial judge explained that the jury could return a guilty verdict based on the circumstantial evidence if guilt was the only rational inference they could draw from the evidence. This much was correct, and as explained in R. v. Cooper, 1977 CanLII 11 (SCC), [1978] 1 S.C.R. 860, it was unnecessary to say more. Unfortunately, the trial judge did. The judge twice proceeded to add that the accused was entitled to an acquittal if the jury found that there was an “equally” rational inference suggesting innocence. On a third occasion, the trial judge stated that “[i]f there is a second inference that’s as reasonable, you will not be able to base a verdict of guilt on circumstantial evidence” (emphasis added). Having chosen to elaborate on the basic principle, the trial judge should have said that if there was any other (as opposed to an equally) rational inference arising from the proven facts that did not point to guilt, the jury would be bound to acquit. The trial judge’s instruction, read as a whole, might well have been understood to suggest that a reasonable inference not pointing to guilt will be insufficient for an acquittal if it is not as reasonable as the inference that the accused is guilty. As Doyon J.A. noted, such a standard would come “dangerously close” to a balance of probabilities (2008 QCCA 824, 2008 QCCA 825, 237 C.C.C. (3d) 374, at para. 43).

[80] Following the trial judge’s initial charge, the jury returned with the following question:

Can we please get a legal definition of reasonable doubt? What things can it be based upon, for example, feelings, intuitions, et cetera? If mostly based on circumstantial evidence, where do we draw the line? Please advise.

In his response to this specific question from the jury, the trial judge repeated three more times the incorrect component of this instruction, using the term “equally”. Because the jury was likely to have paid particular attention to the answer to their question, this would have amplified the misdirection (see R. v. Brydon, 1995 CanLII 48 (SCC), [1995] 4 S.C.R. 253, at para. 16).

[81] Pursuant to an objection by defence counsel, the trial judge gave additional instructions, finally using the correct language. However, he did not draw the jury’s attention to the repeated error in his previous instructions. On the contrary, he reiterated a statement that he had made to the jury several times before, merely omitting, on this occasion, the incorrect term “equally”. This could not possibly have dispelled the confusion.

[82] We agree with our colleague that jury instructions should be read as a whole. However, this reading must include the correct and the incorrect directions. The defective statements are not eclipsed by a correct statement of the law elsewhere in the charge. If an incorrect instruction has been given, the trial judge must advert to it and correct it clearly. Otherwise, the unidentified incorrect statements may linger in the juror’s minds, creating confusion as to the state of the law and obliging the jury to decide on their own which of the competing expositions of the law to apply to the facts. In this sense, the jurors would be left to instruct themselves as to the law they were required to apply in reaching their verdict.

[83] In order to ensure that a misdirection to the jury has been successfully rectified, a trial judge will generally have to repeat the direction he or she has given, acknowledge that it was incorrect, tell the jury to put it out of their minds and then tell them in plain and simple terms what the law is: R. v. Moon, [1969] 3 All E.R. 803 (C.A.), cited in R. v. Seguin (1979), 1979 CanLII 2948 (ON CA), 45 C.C.C. (2d) 498 (Ont. C.A.), at p. 503. See also R. v. Boyd (1999), 1999 CanLII 2870 (ON CA), 118 O.A.C. 85, at para. 29; R. v. Carpenter (1993), 1993 CanLII 8591 (ON CA), 14 O.R. (3d) 641 (C.A.), at p. 647; R. v. Cavanagh (1976), 1976 CanLII 649 (ON CA), 15 O.R. (2d) 173 (C.A.), at p. 182; and Shapiro v. Wilkinson, 1943 CanLII 14 (ON CA), [1943] O.R. 806 (C.A.), at pp. 811-12.

[84] The second deficiency in the instructions on the burden of proof arose in the trial judge’s response to the same question from the jury:

So, when you ask in your question, is it feelings, is it intuitions, I would answer no to that question. Whether you find there is a reasonable doubt, or whether you find there is not a reasonable doubt, you should individually be in a position to be able to explain your position. You should be able to say, for instance, “Listen, I have a reasonable doubt on this essential element, and here’s why I have a reasonable doubt. I can’t rely on Witness A and B. I don’t find Witness A and B credible. And here’s why I don’t find Witness A and B credible, because there’s contradictions, or they’re not confirmed in the evidence, and I don’t accept that Exhibits X, Y and Z help me out on this particular issue,” so that you’re able as a judge to explain why you are supporting a particular position. [Emphasis added; A.R., at pp. 211-12.]

[85] This statement raises the following issue: may a judge instruct a jury that a reasonable doubt is a doubt for which one can give a reason or an explanation? In R. v. Lifchus, 1997 CanLII 319 (SCC), [1997] 3 S.C.R. 320, at paras. 28-30, Cory J. objected to such an instruction on the ground that it might dissuade an “inarticulate juror” from concluding that a doubt is reasonable. Furthermore, he explained, there are certain doubts that, while reasonable, “are simply incapable of articulation”. He concluded that it is “certainly not essential” to instruct a jury that they should be able to explain their doubts. Nevertheless, this Court later ruled that such an instruction would not necessarily constitute reversible error: R. v. Russell, 2000 SCC 55, [2000] 2 S.C.R. 731, and R. v. Beauchamp, 2000 SCC 54, [2000] 2 S.C.R. 720.

[86] In this case, however, as the underlined portion of the judge’s response indicates, the instructions went one step further than in Russell and Beauchamp, requiring the jury to be able to identify quite a specific reason for their doubt. Not only were they directed that they must be able to explain, that they would acquit because they did not believe a particular witness, or were not persuaded by a particular piece of evidence, but they would also have to explain what it was about that testimony or item of evidence that did not convince them. Such precision is generally not even required of a trial judge. As Bastarache and Abella JJ. explained in R. v. Gagnon, 2006 SCC 17, [2006] 1 S.C.R. 621, at para. 20:

Assessing credibility is not a science. It is very difficult for a trial judge to articulate with precision the complex intermingling of impressions that emerge after watching and listening to witnesses and attempting to reconcile the various versions of events.

It cannot be correct to hold a jury to so high a standard. As Cory J. stated in Lifchus, at para. 29, “[a] juror should not be made to feel that the overall, perhaps intangible, effect of a witness’s demeanour cannot be taken into consideration in the assessment of credibility.”

[87] Whether an error in a recharge warrants a new trial is determined according to the standard set out in Brydon, at para. 19:

[W]hether, after placing the inconsistency or error in the context of the charge as a whole, there is a reasonable possibility that the jury might have been misled by those instructions into either applying a standard of proof less than proof beyond a reasonable doubt or improperly applying the burden of proof or reasonable doubt standard in arriving at their verdict.

In light of either of the uncorrected deficiencies in the charge on the burden of proof, and considering their combined effect in particular, it cannot be said that there is no reasonable possibility the jury was misled in this case. As mentioned earlier, the misdirection was repeated, notably in the judge’s response to a specific question from the jury, and the purported correction of only one of the two mistakes was insufficient. The error was serious. It may well have had significant consequences. The curative proviso can therefore not be applied. For these reasons, we agree with the majority in the Court of Appeal with respect to the jury instructions on the burden of proof, and would order a new trial on that basis.

B. The Out of Court Statement of Poirier

[88] With respect for the contrary opinion of Charron J., we conclude that the trial judge erred as well in admitting the victim’s impugned out-of-court statement. Moreover, the curative proviso under s. 686(1)(b)(iii) of the Criminal Code, R.S.C. 1985, c. C-46, cannot remedy this error either.

[89] We begin by recalling that hearsay evidence, like any other, must be relevant to be admissible (R. v. Blackman, 2008 SCC 37, [2008] 2 S.C.R. 298, at para. 29). It must tend to “increase or diminish the probability of the existence of a fact in issue” (R. v. Arp, 1998 CanLII 769 (SCC), [1998] 3 S.C.R. 339, at para. 38). Under the rule against hearsay, the presumption is that out-of-court statements are inadmissible to prove the truth of their contents (Blackman, at para. 33), absent an exception. One should first look to the traditional exceptions (R. v. Mapara, 2005 SCC 23, [2005] 1 S.C.R. 358), and then to the principled exception (R. v. Khan, 1990 CanLII 77 (SCC), [1990] 2 S.C.R. 531). The traditional exceptions are presumptively valid, though they are still subject to modification in light of the principled analysis of necessity and reliability (R. v. Starr, 2000 SCC 40, [2000] 2 S.C.R. 144). The admissibility decision is also informed by the judge’s residual discretion to exclude evidence where its probative value is outweighed by the potential prejudicial effect (R. v. Hawkins, 1996 CanLII 154 (SCC), [1996] 3 S.C.R. 1043, at para. 85; R. v. Smith, 1992 CanLII 79 (SCC), [1992] 2 S.C.R. 915, at p. 937). On appeal, the trial judge’s decision is entitled to deference, and should only be disturbed where there is an error in principle or the decision is not supported by the evidence (Blackman, at para. 52).

[90] In our respectful view, the trial judge in this case committed an error in principle.

[91] Jennifer Williams testified that about two weeks before his killing, her boyfriend, the victim Denis Poirier, stated to her that “—i—f anything happens to me it’s your cousin’s family.” She explained that in light of their previous conversations, his statement could only have referred to the respondent, John Griffin.

[92] There are four purposes for which the statement might conceivably have been admitted: first, to prove Griffin’s state of mind; second, to eliminate others who may have wished Poirier harm; third, to establish Poirier’s state of mind; and fourth, to bolster other circumstantial evidence of motive. It is instructive to reconsider these four theories, and why they all ultimately fail.

[93] The first possible theory is that Poirier’s statement was admissible to support an inference that Griffin had a motive to harm the victim. In this case, motive was relevant to the issue of identification. No traditional exception accommodates this first theory: a principled analysis is therefore required to determine whether the statement was sufficiently reliable and necessary to address the traditional hearsay dangers.

[94] The first criterion, necessity, is easily established by the fact that the declarant is dead: see R. v. Khelawon, 2006 SCC 57, [2006] 2 S.C.R. 787.

[95] With respect to reliability, the problem with hearsay statements is that there is no way to inquire into the declarant’s “perception, memory, narration, or sincerity” (Law Reform Commission of Canada, Report on Evidence (1975), at p. 69 (emphasis added)). In other words:

[The declarant’s] statement about the fact might be false because he misperceived it or did not remember it correctly, or he may have misled the person to whom it was made because he used words not commonly used, or he may simply have lied about it. [Ibid.]

[96] In the circumstances of this case, there was little reason to believe that the declarant’s memory would be a problem. And the trial judge noted that Williams would explain that in light of her previous conversations with her boyfriend, his phrase “your cousin’s family” could only have referred to Griffin, so the hearsay danger associated with narration was also addressed. Similarly, the circumstances in which the statement was made alleviated concerns about sincerity: as the trial judge observed, the statement was made to the declarant’s girlfriend, it was responsive to prior dialogue, and there was no evidence of a motivation to lie as there was in Smith and Starr. These considerations formed the basis of the trial judge’s decision that the statement could be admitted into evidence.

[97] We agree that there is little concern as to whether Poirier was sincere in his belief that Griffin might do him harm. Sincerity, however, is not the principal problem with Poirier’s statement. Rather, it is perception: we simply do not know the basis for the victim’s statement that the respondent wished him harm, or whether his perception was accurate. In Starr, the impugned out-of-court statement was ruled inadmissible for precisely this reason. Speaking for the majority, Iacobucci J. refused to admit the hearsay statement to prove the intentions of a third party:

[T]he central concern with hearsay is the inability of the trier of fact to test the reliability of the declarant’s assertion. When the statement is tendered to prove the intentions of a third party, this danger is multiplied. If a declarant makes a statement about the intentions of a third party, there are three possible bases for this statement: first, it could be based on a prior conversation with the accused; second, it could be based on a prior conversation with a fourth party, who indicated the third party’s intentions to the declarant; or third, it could be based on pure speculation on the part of the declarant. Under the first scenario, the statement is double hearsay. Since each level of double hearsay must fall within an exception, or be admissible under the principled approach, the mere fact that the declarant is making a statement of present intention is insufficient to render it admissible. The second level of hearsay must also be admissible.

The other two scenarios also clearly require exclusion. If the statement about joint acts is based on a conversation with a fourth party, then the statement is triple hearsay, or worse. If, on the other hand, it is based on pure speculation, then it clearly is unreliable . . . . [paras. 172-73]

Exactly the same problems arise in our case.

[98] Beyond the concerns discussed in Starr, an additional problem with the statement in question — “[i]f anything happens to me it’s your cousin’s family” — is that it is a statement of opinion. As Major J. observed in R. v. D.D., 2000 SCC 43, [2000] 2 S.C.R. 275, at para. 49:

A basic tenet of our law is that the usual witness may not give opinion evidence, but testify only to facts within his knowledge, observation and experience. This is a commendable principle since it is the task of the fact finder, whether a jury or judge alone, to decide what secondary inferences are to be drawn from the facts proved.

Here, it is not clear what “knowledge, observation and experience” — if any — led the victim to form his opinion as to Griffin’s future intentions. And no hearsay exception overcomes the fact that it is opinion evidence. As Professors David Paciocco and Lee Stuesser explain, “if the witness could not testify as to the statement in court, it is not rendered admissible simply because it was made out of court” (The Law of Evidence (5th ed. 2008), at p. 135).

[99] Why the victim believed Griffin wished him harm is purely a matter of speculation. Evidence was tendered to show that a drug deal between Poirier and Griffin had gone awry and that Griffin was owed $100,000; that Griffin was seeking Poirier out and Poirier was hiding from him; that Poirier’s friend Joe Besso was missing; and that Poirier had said that he was “scared for his life”. But it is also possible that Poirier’s opinion was simply the product of a paranoid imagination. Thus, if we are asked to provide the basis for the declarant’s opinion by analysing the other evidence, it would be dangerous to form any conclusions that we could not already derive from that other evidence. In short, the statement of opinion adds nothing, and there is a substantial risk that it was accorded undue weight: it lacked probative value, and it was highly prejudicial. The reasoning of Heeney J. in R. v. Cowell, [2002] O.J. No. 4783 (QL) (S.C.J.), at paras. 71-74, illustrates this concern well:

The witness said: “she was afraid if she left him, he’d hurt himself”. The Crown seeks to admit this under the state of mind exception.

Assuming that this statement is admissible as evidence of the state of mind of the victim, I am satisfied that its prejudicial impact outweighs its probative value. The difficulty with the statement is that it expresses the concern of the victim that the accused would hurt himself, without articulating any basis for that belief. That belief might be based on mere speculation. The Crown argues that there is other evidence that will provide support for the conclusion that the accused was planning to commit suicide. He borrowed money to pay off the substantial debt he owed to his parents, for example, showing he was getting his financial affairs in order. I have already mentioned the evidence found after the fact that is consistent with a suicide attempt having been made or, at least, contemplated.

That evidence can be considered by the jury in deciding whether the accused did, indeed, intend to harm himself, but the baldly stated opinion of the victim that the accused might hurt himself if she left adds nothing of value to the evidence. It is the opinion of the jury, not the absent witness, that counts, particularly when the victim’s opinion might have been based on nothing more than speculation. Had she been alive to testify, she would not have been permitted to state her opinion as to what the accused might do if she left him, without articulating the factual basis for feeling that way.

While the jury could be permitted to use this evidence as indicating the state of mind of the victim at that time, there is a grave danger that the jury might misuse this evidence to infer that there was, indeed, good reason to believe the accused was suicidal, despite any limiting instruction I might give. That would be highly prejudicial to the accused. Since the state of mind of the victim one month before her death is only marginally probative, I conclude that the prejudicial impact of this evidence outweighs its probative value.

Here, too, the out-of-court statement was of marginal probative value at best, and its inclusion plainly occasioned a substantial risk of prejudice to the accused. As in Cowell, the impugned statement was certainly not admissible on the first theory, that is to establish Griffin’s motive and thereby identify him as the murderer.

[100] The second possible theory is the one the trial judge accepted: that the statement was “relevant to rebut the notion that others would also have had a motive to kill [Poirier] at the beginning of 2003— (2005 CanLII 5629, at para. 32). Having admitted the statement, the trial judge instructed the jury not to use it to infer motive on the part of Griffin, but explained that it could be used to “eliminate other potential people who would want to do him harm, as far as he was concerned”. This would have deeply undermined the respondents’ principal defence.

[101] The majority of the Court of Appeal ruled that the limiting instruction was inadequate, because to suggest that other people did not mean the victim harm could only lead the jury to the conclusion that Griffin was the only person who did — which was precisely the inference that was prohibited. Doyon J.A. therefore concluded that the jury could not possibly have applied the distinction they were asked to draw. Were we to conclude that the statement was admissible, we would agree with Doyon J.A. that the limiting instruction was inadequate.

[102] However, the more fundamental problem is that the statement could not possibly support an inference that others did not wish the victim harm.

[103] The second theory fails for the same reason as the first: we are asked to use Poirier’s opinion to infer the state of mind of third parties, without having any idea of the basis for his conclusions. And using the statement to infer the state of mind of “others” is even more problematic than using it to infer the state of mind of Griffin, for four reasons. First, the “others” contemplated are not even identified. Second, the principal meaning of the statement seems to be that Poirier believed Griffin meant him harm, and it is not clear that he also thought that nobody else did. Third, the most one could infer is that the victim lacked any reason to believe that others wished him harm, which is even less helpful than an inference that he did have some reason to believe Griffin wished him harm. Fourth, and finally, if a hearsay statement can be permitted to help establish anybody’s state of mind, it should be that of the accused, because the accused is available to testify or bring evidence to contradict such an inference in court. Indeed this is one of the principal reasons that an exception to the rule against hearsay is made for party admissions (see J. Sopinka, S. N. Lederman and A. W. Bryant, The Law of Evidence in Canada (2nd ed. 1999), at p. 287, —6.292). Unidentified third parties, needless to say, do not have the same ability. For all of these reasons, the statement was not admissible to “eliminate” other potential suspects.

[104] The third conceivable use of the statement is to establish that Poirier was afraid of Griffin — in other words, to establish the state of mind of the victim. At first blush, this use would seem to fall within the scope of the “state of mind” exception to the hearsay rule. However, it is necessary to recall the basic principle: that evidence should only be admitted where it is relevant to an issue in the case. This principle applies equally to our concern. As Sopinka, Lederman and Bryant explain, “[i]f the mental state of the declarant is directly in issue at trial, then statements of his or her mental state are generally admissible in proof of the fact” (p. 256, —6.225). On the other hand, the state of mind of the victim is irrelevant unless it tends to support a permissible inference regarding a relevant fact.

[105] Charron J. suggests (at para. 59) that Poirier’s fear of Griffin was relevant to Griffin’s motive, and in turn to the issue of identification. We agree that the statement “[i]f anything happens to me it’s your cousin’s family” can be understood not only as speculation, but also as an expression of fear, which is a state of mind. It could also be inferred that Poirier believed that Griffin had a motive to harm him. Again, however, we cannot be certain what basis Poirier may have had for such a belief, and therefore cannot rely upon this statement. With respect for those of a contrary opinion, we do not believe the statement is admissible under this theory. In this case, the victim’s state of mind is irrelevant on its own, and it is impermissible to use it to infer Griffin’s motive.

[106] This brings us to the fourth possible use of the statement: was it admissible to support other circumstantial evidence in the case? There was evidence suggesting that Poirier owed Griffin $100,000, that Poirier was unable to pay, and that Poirier was hiding from Griffin. This evidence related to the critical issue of motive. Poirier’s statement to Williams — “[i]f anything happens to me it’s your cousin’s family” — does tend to support these other facts. It was consistent with Poirier’s situation that he should have said what he did to his girlfriend. In the language of Arp, it “tend[s] to —increase . . . the probability of the existence of a fact in issue—” (para. 38). Under this theory, the statement is tendered not for the truth of its contents, but for the mere fact that it was made. It thus avoids the problem identified above, that the basis for the statement is not known.

[107] Nevertheless, the statement could not be admitted under this theory either, because its probative value for this purpose is greatly outweighed by its potential prejudicial effect. As the respondent Griffin states, Poirier’s statement “begged to be used by the jury for the prohibited purpose” of inferring Griffin’s intent (Factum, at para. 82). When the jury hears “[i]f anything happens to me it’s your cousin’s family”, it is simply too tempting for the jury to reason that —something happened: it must have been his cousin’s family—. The statement could assume an almost superstitious significance.

[108] This great risk of prejudice could not be attenuated by a limiting instruction. It is difficult to justify admitting the statement for a marginally probative and tangential purpose while insisting that the jury not use it in the most obvious and prejudicial way possible. Numerous cases indicate that juries should be trusted to use evidence as directed: R. v. Corbett, 1988 CanLII 80 (SCC), [1988] 1 S.C.R. 670, at p. 692; R. v. B. (F.F.), 1993 CanLII 167 (SCC), [1993] 1 S.C.R. 697; R. v. W. (D.), 1991 CanLII 93 (SCC), [1991] 1 S.C.R. 742; R. v. Spence, 2005 SCC 71, [2005] 3 S.C.R. 458, at para. 22; and Smith. But there does come a point, as in this case, where one must heed Cardozo J.’s warning: “The reverberating clang of those accusatory words would drown all weaker sounds. It is for ordinary minds, and not for psychoanalysts, that our rules of evidence are framed” (Shepard v. United States, 290 U.S. 96 (1933), at p. 104).

[109] We conclude therefore that it was wrong to have admitted the statement in the first place. This amounted to an error in principle, and deference cannot be accorded to the judge’s decision.

[110] We would also decline to apply the curative proviso of s. 686(1)(b)(iii) in respect of this error. Neither of the possibilities discussed in R. v. Khan, 2001 SCC 86, [2001] 3 S.C.R. 823, at para. 26, apply here. On the one hand, the decision to admit the statement cannot be characterized as “harmless”: Poirier’s fingering of Griffin from beyond the grave, so to speak, could only have had tremendous impact, despite the limiting instruction. Nor can the evidence against the accused be described as “so overwhelming that . . . there was no substantial wrong or miscarriage of justice”. The Crown’s case rested in significant measure on circumstantial evidence and on the credibility of certain witnesses, and so we are not persuaded that no properly charged jury could have come to another verdict.

III. Conclusion

[111] For the foregoing reasons, and with respect for those who are of a contrary opinion, we would dismiss the Crown’s appeals. We would affirm the order of the Court of Appeal for a new trial for both John Griffin and his alleged accomplice, Earl Roy Harris.

Appeals allowed, LeBel and Fish JJ. dissenting.

Solicitor for the appellant: Poursuites criminelles et pénales du Québec, Montréal.

Solicitors for the respondent Griffin: Filteau Belleau, Montréal.