CITATION: R. v. Quidley, 2008 ONCA 501
DATE: 20080623
DOCKET: C44210
COURT OF APPEAL FOR ONTARIO
O’CONNOR A.C.J.O., BLAIR and LANG JJ.A
BETWEEN:
HER MAJESTY THE QUEEN
Respondent
and
ALLAN QUIDLEY
Appellant
James Lockyer for the appellant
Kevin Wilson for the respondent
HEARD: March 27, 2008
On appeal from the convictions entered by Justice F. Dawson, sitting with a jury, on June 16, 2005 and from the sentence imposed on September 15, 2005.
R.A. BLAIR J.A.:
Overview
[1] Mr. Quidley seeks to set aside his convictions on serious drug-related offences which formed part of a larger conspiracy to import cocaine, hash oil, hashish, and marijuana through Pearson International Airport on flights from Jamaica. He received a global sentence of fifteen years imprisonment. In separate proceedings, at least six others were also convicted of various offences arising out of the same circumstances.
[2] The appellant worked as a groomer for Air Canada. He cleaned the interior of planes when they arrived at the airport. The Crown alleges that he and other employees used this position with the airline to retrieve drugs hidden on planes by couriers who had boarded in Jamaica. The investigation into this drug conspiracy – Project O’Booker – covered a period of approximately one year and produced a large number of intercepted phone calls said to have taken place between the appellant and others. Many of the conversations were conducted in Jamaican Patois and had to be translated into English. These phone calls were central to the case against the appellant. The accuracy of the voice identification and the accuracy of the translations from Patois to English were thus key components of the trial. It is the way in which the trial judge dealt with these issues in his charge to the jury that is at the heart of the appeal.
[3] Mr. Lockyer also raised a Vetrovec issue in relation to a key Crown witness, Jeffrey Cahill, who was also an Air Canada employee and who testified about the nature and scope of the conspiracy and the role of Air Canada groomers in it. He was permitted as well to provide voice identification in some instances. We did not call on the Crown with respect to the Vetrovec issue and, on the view I take of the appeal, it is not necessary to deal with that issue directly.
[4] I am satisfied that the appeal must be allowed on the first ground raised, namely that the trial judge misdirected the jury on the burden of proof in relation to the voice identification evidence and the accuracy of the translations from Patois into English. This error could not have been overcome by the otherwise flawless charge with respect to the burden of proof and reasonable doubt. My reasons for arriving at this conclusion follow.
Facts
[5] Project O’Booker began in November 1999 and lasted for about a year. The appellant was arrested on November 29, 2000 and ultimately convicted of:
a) conspiracy to import cocaine into Canada between November 1, 1999 and November 29, 2000;
b) conspiracy to import cannabis resin into Canada between November 1, 1999 and November 29, 2000;
c) trafficking in cocaine on September 26, 2000;
d) importing cocaine into Canada on October 22, 2000; and
e) importing cannabis resin into Canada on October 22, 2000.
In each case, telephone interceptions of conversations purporting to involve the appellant were key to establishing the conviction. Although large quantities of drugs seized over the course of Project O’Booker were alleged to be linked to the conspiracy, none of them were ever directly linked to the appellant.
[6] A total of forty-nine kilograms of cocaine, one hundred and three kilograms of hashish, and fifty-five kilograms of hash oil were seized in the course of the investigation. An expert witness testified for the Crown as to the value of these drugs and this evidence was not challenged by the defence. The cocaine seized would have generated over $1.3 million if sold by the kilogram, $2.07 to $3.11 million if sold by the ounce, and upwards of $12 million if sold by the gram. The hashish would have sold for between $1.03 and $5.12 million, and the hash oil could have generated $734,000 or more. It was not an inconsequential criminal operation that was foiled.
[7] As noted above, the Crown alleged that the appellant and his co-conspirators used their position as groomers and employees at Air Canada to retrieve drugs hidden on airplanes flying from Jamaica to Pearson International Airport. Couriers, boarding in Jamaica, secreted the drugs behind easily removable wall panels in the passenger seating area of the planes or in the ceilings and other areas of the washrooms. The appellant and his co-conspirators are alleged to have picked up the drugs during the process of cleaning the planes and to have sold them to other individuals in Canada.
[8] Little of Project O’Booker’s surveillance evidence – except for the September 26, 2000 incident involving trafficking in cocaine – connected the appellant to the crimes. However, he was alleged to have been a participant in a large number of intercepted wiretap conversations concerning the conspiracies. The wiretap connection was central to the Crown’s case on all convictions.
[9] A total of eighty-six intercepted telephone calls were entered into evidence against the appellant. Many of them were conducted in Jamaican Patois and required English translation. While the defence admitted the time, date, and place of the calls, the voice identification and the accuracy of the translations from Patois to English were hotly contested issues at trial. In this respect, the evidence of two witnesses – Jeffrey Cahill and P.C. Williams – was particularly important.
[10] The first witness was Jeffrey Cahill. He was permitted to give voice identification evidence. Mr. Cahill was a former Air Canada employee and a major participant in the conspiracy. Shortly before the appellant’s trial, he agreed to testify for the Crown against the other alleged members of the conspiracy. In exchange the Crown agreed to – and did – stay all charges against Mr. Cahill, including conspiring to import drugs, importing drugs, and certain firearms charges arising from his arrest. The agreement between Mr. Cahill and the Crown was filed as an exhibit at trial.
[11] Apart from his evidence about the nature and scope of the conspiracy and the role of Air Canada groomers in it, Mr. Cahill listened to the telephone interceptions and testified as to the identity of many of the participants, including the appellant. Indeed, the Crown relied upon Mr. Cahill as their sole evidence of voice identification. When the interceptions were played to him in the witness box, Mr. Cahill had the transcripts of those conversations in front of him with the alleged speakers already identified on the page. A similar process had been followed prior to the date he gave his evidence when he met with Crown counsel and a police officer to review the tapes. Mr. Cahill testified that in almost all of the instances in which the appellant was purported by the transcript identification to be speaking in the conversations, he was able to identify the appellant’s voice. The defence challenged these identifications, arguing that Mr. Cahill had motive to shape his testimony in favour of the Crown and that the resulting identifications were particularly dangerous since they lay at the very heart of the Crown’s case against the accused.
[12] The second witness was P.C. Williams of the Peel Regional Police. P.C. Williams was permitted to provide opinion evidence on the accuracy of the Patois translations and testified that the translations were accurate. The trial judge ruled that the jury should also hear the interceptions themselves to listen for voice intonation and emotion. The trial judge instructed the jurors that in the case of the English portions of the recordings, the evidence was the recordings themselves, and in the case of the Patois portions, the transcripts were the primary evidence – so long as they were prepared to rely upon the evidence of P.C. Williams.
[13] That the intercepted conversations were critical to the case is apparent from the position taken by the defence at trial. The trial judge heard pre-charge submissions from counsel and asked each counsel to provide him with a summary of their positions so that he could put them accurately to the jury. Counsel did so, and in his charge to the jury the trial judge gave the summary of the defence as provided to him. The defence raised four points:
a) The Crown had not proven the voice identification in the phone calls where it was alleging the appellant was the speaker;
b) The jury should use the translated transcripts with caution because of the frailties in P.C. Williams’ knowledge of Patois;
c) Even if the jury believed the Crown had proven voice identification and the accuracy of the translation, the jury should carefully consider what it was the parties to the intercepted conversations were speaking about, i.e. that the subject matter of the calls was not clear and that too many assumptions had to be made to come to the conclusion that the appellant was definitely talking about drugs and airplanes; and
d) Even if the jury believed the appellant was definitely talking about drugs and airplanes, it was more probable than not that the intercepted conversations had to do with a plot to steal the drugs from Mr. Cahill rather than with a conspiracy to import them into Canada.
[14] Thus, three out of four of the defence’s main points turned on voice identification and the accuracy of the translations from Patois to English.
Analysis
[15] On behalf of the appellant, Mr. Lockyer argues that the trial judge misdirected the jurors on the burden of proof by instructing them that in the vital areas of voice identification and the accuracy of the translations they need only be satisfied on a balance of probabilities. In so doing, Mr. Lockyer submits, the trial judge left it open to the jury, in the circumstances of this case, to convict on the basis of that standard of proof.
[16] The trial judge instructed the jury – on several occasions – that both voice identification and the accuracy of the translations from Patois to English were matters that did not have to be established beyond a reasonable doubt because they were not essential elements of the offence, and that the jurors need only be satisfied with respect to those matters on a balance of probabilities. For example, in his jury instructions on the voice identification evidence the trial judge said:
Voice identification is not something that must be established beyond a reasonable doubt because it is not an essential element of the offence. However, for any calls where the identity of the speaker matters, you should be satisfied that the speaker has been identified to a probability. In other words, that the evidence establishes that it is more probable than not that the speaker is a particular identified individual.
[After reviewing some of the evidence, the trial judge returned to this theme.]
In any event, it is you that must be satisfied of the voice identification on the balance of probabilities – more probable than not – in any case where voice identification matters. You make that determination from a consideration of all of the evidence. Frailties in the voice identification go to the weight you are prepared to give the evidence in determining if someone’s voice is identified.
Now you may have noticed that I spoke of voice identification, … ‘where it matters’. It will always matter in the case of Mr. Quidley. Before you can attribute a statement from the wiretap evidence to Mr. Quidley, you must be satisfied that he is the speaker who said it. That is on the balance of probabilities, as I said. Before you can look at the contents of the telephone call as evidence admissible directly against Mr. Quidley, you must be satisfied that he is a participant in the conversation. [Emphasis added.]
[17] With respect to the use of the balance of probabilities standard when determining the accuracy of the translations, the trial judge instructed the jury as follows:
[Defence counsel] suggested in her address to you – and this is no criticism of her, I just want to make sure you understand something properly – that you need to be completely satisfied as to the translation. I want you to understand that the accuracy of the translation is not something that must be established beyond a reasonable doubt. That is because it is not an element of the offence. Before relying on the translation of a call or calls it is sufficient if you are satisfied that the translation is probably accurate. That is, that it is more probable than not that the translation is accurate. Aside from that, any deficiencies that you may find in the translation process are matters which go to the issue of the weight you will give the evidence.
So with respect to the transcripts, for English they may be used to the extent you find they are accurate. If there are differences, rely on the recording. For Patois, the transcripts are the primary evidence, provided you are prepared to accept and rely upon the opinion evidence of Constable Williams as to the probability of the accuracy of the translation. [Emphasis added.]
[18] During deliberations the jury posed a question to the trial judge: “Is the transcript agreed to be an accurate representation of the wiretaps?” The trial judge was concerned about this question, as he thought it might have meant that the jury had fundamentally misunderstood his earlier instructions. After discussions with counsel, he re-charged the jury in essentially the same terms as he had previously done. He resisted the Crown’s suggestion at that time to provide the jury with his written charge, but ultimately he did so.
[19] On behalf of the Crown, Mr. Wilson does not seek to support the trial judge’s use of the balance of probabilities standard in relation to the charge respecting voice identification and the accuracy of the translation. Indeed, he concedes there is nothing in the case law to justify it. He submits, however, that the trial judge gave an otherwise solid charge on the burden of proof and reasonable doubt and that upon a review of the whole of the charge, there is no reasonable likelihood that the jury misapprehended the correct standard of proof – namely, proof of guilt on all counts beyond a reasonable doubt: see R. v. Starr (2000), 2000 SCC 40, 147 C.C.C. (3d) 449 at para. 233 (S.C.C.); R. v. Rhee (2001), 2001 SCC 71, 158 C.C.C. (3d) 129 at para. 32 (S.C.C.); R. v. Bisson (1998), 1998 CanLII 810 (SCC), 121 C.C.C. (3d) 449 at para. 9 (S.C.C.); and R. v. Lifchus (1997), 1997 CanLII 319 (SCC), 118 C.C.C. (3d) 1 at para. 41 (S.C.C.).
[20] In this respect, the Crown places considerable emphasis on the portions of the charge where the trial judge stated that deficiencies or frailties in the voice identification or the accuracy of the translations go to the weight of that evidence. The trial judge followed most of his references to the balance of probabilities standard with respect to voice identification and the accuracy of the translation with this instruction. The trial judge used language to this effect:
Aside from that, any deficiencies that you may find in the translation process are matters which go to the issue of the weight you will give to the evidence.
You make that determination [i.e. the determination of voice identification on the balance of probabilities] from a consideration of all of the evidence. Frailties in the voice identification go to the weight you are prepared to give the evidence in determining if someone’s voice is identified.
[21] In addition, in responding orally to the above mentioned question from the jury about the accuracy of the translations, the trial judge told the jury:
As long as you’re able to say from the evidence … that the interpretation is probably accurate, you can take it into account and consider any frailties or whatever you think it has and give it the weight that you determine you should give it in your overall assessment. [Emphasis added.]
[22] This is the closest the trial judge came to directing the jury that, having been satisfied about the voice identification and the accuracy of the translations, they must nonetheless still consider that evidence as part of their overall assessment of the evidence on a beyond a reasonable doubt basis. The Crown fairly acknowledges, however, that the latter instruction was part of the oral re-charge, whereas the original flawed language regarding balance of probabilities was included in the original charge and in the copy of the written charge that was later – also in the midst of deliberations – given to the jurors.
[23] In the end, the Crown submits that the foregoing caveats, coupled with the language of the model Lifchus charge, were sufficient to ensure that the jury did not misapprehend the Crown’s overall burden to prove the appellant’s guilt beyond a reasonable doubt. The Crown points to the following general language, in particular:
It is not enough for you to believe that an accused is probably or likely guilty. In those circumstances you must find the accused not guilty because Crown counsel would have failed to satisfy you of guilt beyond a reasonable doubt. Proof of probable or likely guilt is not proof of guilt beyond a reasonable doubt.
Absolute certainty is a standard of proof that is impossibility high. However, proof beyond a reasonable doubt falls much closer to absolute certainty than to proof of probable or likely guilt.
If, at the end of the case, after considering all of the evidence, you are sure that the accused committed an offence, you should find him guilty of it since you would have been satisfied of guilt of that offence beyond a reasonable doubt. [Emphasis added]
[24] I do not accept the Crown’s submissions in this regard.
[25] In the particular circumstances of this case, it constituted reversible error on the part of the trial judge to charge the jury as he did with respect to the evidence of voice identification and the accuracy of the translations. In my view, there was a very serious likelihood that the jury would be confused and would misapprehend the correct standard of proof.
[26] The voice identification evidence and the evidence of P.C. Williams respecting the accuracy of the translations were pivotal to the case. Without the jury deciding that it was the appellant’s voice they heard on the tapes and that P.C. Williams’ translation of those conversations was accurate, the Crown had no case against the appellant. If the jury decided in the Crown’s favour on those questions, there was every likelihood the appellant would be convicted, as indeed he was. These two areas of evidence – while not essential elements of the offences charged – were sufficiently central to the determination of the case that they were, for practical purposes, almost the same as essential elements on this record.
[27] The trial judge focused on the point that neither the voice identification nor the accuracy of the translations was an essential element of the offence. Accordingly, he instructed the jurors that these matters need not be established beyond a reasonable doubt. There was no error in that direction in itself. However, his decision to instruct the jurors that they must be satisfied on these matters on a balance of probabilities, and the manner in which he put it to them, were problematic. It left the jurors with the impression that, once they had determined on a balance of probabilities (a) that Mr. Quidley was the speaker who said what was said on the wiretap, and (b) that the translation of that conversation was accurate, that was the end of the exercise respecting those pieces of evidence. In effect, the jurors were told to make a finding with respect to them and were then asked to incorporate their probability finding as a fact in their decision on reasonable doubt. The jurors were not told clearly that they must consider the voice identification evidence and the evidence respecting the accuracy of the translation – together with all of the other evidence – in determining whether that evidence, or any of the other evidence, raised a reasonable doubt about the appellant’s guilt.
[28] This danger is apparent from the following passage of the charge, for example:
In any event, it is you that must be satisfied of voice identification on a balance of probabilities – more probable than not – in any case where voice identification matters. You make that determination from a consideration of all of the evidence. Frailties in the voice identification go to the weight you are prepared to give to the evidence in determining if someone’s voice is identified.
Now you may have noticed that I spoke of voice identification … ‘where it matters’. It will always matter in the case of Mr. Quidley. Before you can attribute a statement from the wiretap evidence to Mr. Quidley, you must be satisfied that he is the speaker who said it. That is on a balance of probabilities… [Emphasis added.]
[29] The trial judge’s other instructions on voice identification and the accuracy of the translations leave the same impression. The instructions tell the jurors that once they are satisfied – on a balance of probabilities – that the voice identified is the appellant’s (or, in other instances that the translation from Patois to English is accurate), they may attribute the statement on the wiretap to the appellant (or consider the translation accurate). There is nothing left, then, to consider on a reasonable doubt basis with respect to those issues.
[30] Case law makes it clear (a) that jurors are not to address the evidence piecemeal and seek to apply the criminal standard to separate pieces of evidence before doing so to the evidence as a whole (see R. v. Morin (1988), 1988 CanLII 8 (SCC), 44 C.C.C. (3d) 193 at 210-211 (S.C.C.)), and (b) that jurors are not to examine evidence in two separate stages, first sifting out facts they find to be true and reliable and then, in a second stage, using only those facts to reach their verdict (see R. v. Miller (1991), 1991 CanLII 2704 (ON CA), 68 C.C.C. (3d) 517 at para. 70 (Ont. C.A.)).
[31] In Morin at 210-211 Sopinka J. said:
The argument in favour of a two-stage application of the criminal standard has superficial appeal in theory but in my respectful opinion is wrong in principle and unworkable in practice. In principle, it is wrong because the function of a standard of proof is not the weighing of individual items of evidence but the determination of ultimate issues.
The concern which proponents of the two-stage process express is, that facts which are doubtful will be used to establish guilt. The answer to this concern is that a chain is only as strong as its weakest link. If facts which are essential to a finding of guilt are still doubtful notwithstanding the support of other facts, this will produce a doubt in the mind of the jury that guilt has been proved beyond a reasonable doubt.
I conclude from the foregoing that the facts are for the jury to determine subject to an instruction by the trial judge as to the law. While the charge may and often does include many helpful tips on the weighing of evidence such as observing demeanour, taking into account the interest of the witness and so forth, the law lays down only one basic requirement: during the process of deliberation the jury or other trier of fact must consider the evidence as a whole and determine whether guilt is established by the prosecution beyond a reasonable doubt. This of necessity requires that each element of the offence or issue be proved beyond a reasonable doubt. Beyond this injunction it is for the trier of fact to determine how to proceed….
The reason we have juries is so that lay persons and not lawyers decide the facts. To inject into the process artificial legal rules with respect to the natural human activity of deliberation and decision would tend to detract from the value of the jury system. Accordingly, it is wrong for a trial judge to lay down additional rules for the weighing of the evidence. Indeed, it is unwise to attempt to elaborate on the basic requirement referred to above. I would make two exceptions. The jury should be told that the facts are not to be examined separately and in isolation with reference to the criminal standard. This instruction is a necessary corollary to the basic rule referred to above. Without it there is some danger that a jury might conclude that the requirement that each issue or element of the offence be proved beyond a reasonable doubt demands that individual items of evidence be so proved. [Emphasis added.]
[32] In Miller at para. 70 this Court stated:
In the present case, the impugned direction did not instruct the jury to apply proof beyond a reasonable doubt in two stages. In our opinion, however, it was a misdirection to instruct the jury to examine the evidence in a first stage, to eliminate all evidence except that which the jurors accepted as true and reliable (a lower standard than proof beyond a reasonable doubt), and then to consider only the residual in arriving at their verdict. This involved the injection of artificial rules for the jury and constituted “an intrusion into the province of the jury”. We believe it is wrong and prejudicial to confine the jury, in their initial findings of fact, to evidence accepted as true and reliable; evidence which is neither rejected nor accepted should survive to the final stage of the jury’s determination on the crucial application of reasonable doubt.
[33] Although the specific circumstances prevailing in Morin or Miller do not arise in this case, the fundamental danger underlying the reasons in those cases is inherent in the trial judge’s instructions to the jury. Those cases stand for the proposition that jurors are not to apply the standard of proof to individual pieces of evidence on a piecemeal basis and then use only the evidence that meets that standard to arrive at their verdict. This is for fear that such an approach will skew the jury’s ultimate determination of the overall application of reasonable doubt. The concern is that this skewing may occur either by jurors rejecting evidence altogether that might be capable of giving rise to a reasonable doubt, or by accepting evidence at an earlier stage and not considering it further.
[34] In this case, one of the concerns is that the jurors may have accepted certain evidence on a balance of probabilities, thus removing any potential concerns about that evidence from their consideration of guilt beyond a reasonable doubt. By instructing the jurors that they should first determine the important issues of voice identification and the accuracy of the translations on a balance of probabilities standard without making it clear that the jurors then had to consider whether on all of the evidence – including the voice identification and translation evidence – a reasonable doubt was raised in favour of the accused, the trial judge effectively did the same thing, in my view. The trial judge should not have referred to the balance of probabilities standard in the way he did.
[35] Telling the jury to apply a lower standard of proof to individual pieces of evidence was an error, particularly, in the circumstances of this case, where voice identification and the accuracy of the translations were so critical to the conviction or acquittal. These two issues were central to three out of four of the lines of defence put to the jury on the appellant’s behalf. Some measure of the importance of the instructions, and of the role played by the wiretap evidence, is the fact that the appellant was acquitted on the only count where wiretap evidence was not a factor.
[36] The risks of leaving the jury with the instructions they received were amplified by a number of other factors in this case, as well. First, the only voice identification evidence came from a Vetrovec witness – Mr. Cahill – who had much to gain from currying Crown favour. The jury was told to address this evidence on the lower standard of balance of probabilities in relation to the one issue (voice identification) where his evidence incriminated the appellant. Moreover, voice identification evidence requires careful consideration itself because of its inherent frailties, much like eyewitness testimony: see R. v. Rowbotham (No. 4), [1977] O.J. No. 1686 at para. 46 (Co. Ct. J.), Borins J. Secondly, the jurors were given copies of the trial judge’s written instructions part way through their deliberations and after they had asked a key question about whether the transcripts were agreed to be an accurate representation of the wiretaps – thus fortifying the earlier misdirection. Finally, the trial judge allowed the Crown to file a chart with the jury – not as evidence but for their assistance only – that listed all of the calls where the appellant had been “identified” as one of the callers, thereby enhancing the likelihood that the crucial question of the appellant’s voice identification would be determined decisively on the balance of probabilities standard.
[37] To conclude, then, I am satisfied that the trial judge misdirected the jury with respect to the burden of proof to be applied in the circumstances of this case.
Disposition
[38] Accordingly, I would allow the appeal and direct a new trial.
[39] In view of the foregoing disposition, it is not necessary to address the sentence appeal, which is dismissed as moot.
“R.A. Blair J.A.”
“I agree D. O’Connor A.C.J.O.”
“I agree S. Lang J.A.”
RELEASED: June 23, 2008