Judicially-raised appeal points

Occasionally, the best point in favour of an appellant may only be noticed by the judges of the appellate court. This occurred in Charles v R (Saint Vincent and the Grenadines) [2007] UKPC 47 (16 July 2007). The Board observed (para 17):

“…It is the duty of an appellate court to advert to any such matter which may appear to it to be significant and possibly determinative of an appeal in favour of the accused.”

This had happened in favour of the appellant’s co-accused, whose appeal was allowed by the Court of Appeal; but it was overlooked when the same Court dealt with the appeal by Mr Charles.

The critical point in this case was a misdirection by the trial judge about the use that the jury may make of the statements that each of the two accused had made to the police. The jury had been told, wrongly, that they could use each statement as evidence against the other accused.

This is a long established and simple point, and it has survived statutory reforms of the law of evidence. For example, in New Zealand, the Evidence Act 2006 (to come into force on 1 August 2007) preserves this rule in s 27(1). The reason for the rule is the obvious one that each statement will be unreliable to the extent that it is relevant in relation to the guilt of another accused. At para 8 of Charles, the Board observed of the co-accused “In the statement he fully admitted complicity in the shooting, but tended, as is not unknown in such cases, to throw more of the blame on his accomplice.”

For the proper form of the direction, the Board recommended as follows (para 16):

“In England and Wales the standard models of directions prepared by the Judicial Studies Board and published in its Bench Books are of great assistance to judges, and any similar venture in other common law jurisdictions is likely to be useful.”

A retrial was ordered in this case because there was no discernable reason for differentiating between the accuseds as far as the significance of the misdirection was concerned.

Some observations were made on identification procedures. Here, the witness (a victim who survived) had had a good opportunity to observe the person who she said was Mr Charles, so any deficiencies that had occurred in relation to the jury being warned as to the need for care were not of critical importance in this appeal. At para 12 the Board summarised the way the use of photographs should be treated:

“Their Lordships consider that some care has to be taken when identification from photographs is carried out, although it is not in itself an improper practice. The rules applicable in England and Wales under Code of Practice D, although not binding, form a reliable basis for good practice. Two basic rules are set out in May & Powles, Criminal Evidence, 5th ed (2004), para 14-35:

'(1) The police may show a witness photographs in order to identify a suspect.
'(2) Once a man has been arrested, and there is therefore an opportunity that he can be identified in person, photographs should not be shown to witnesses before an identification parade.'

“As the learned authors point out, when the police are looking for a culprit, the showing of photographs to witnesses may be essential: indeed, it may be the only way in which the culprit can be identified. Once he has been picked out and is available to take part in an identification parade, photographs should not be shown to witnesses. They should instead be asked to attend an identification parade, as should also the witness or witnesses who picked the suspect out from photographs. In relation to the latter, the procedure set out in the headnote to R v Lamb (1980) 71 Cr App R 198 should be followed, viz, the defendant's advisers should be informed of the showing of the photographs and the decision left to them whether to refer to that at trial. If they do so decide, the photographs should not be shown to the jury, and they should be warned of the consequence that the reliability of the identification is likely to be decreased.”

Another topic mentioned in this case is the Mushtaq direction (blogged here 22 April 2005) concerning the need for the jury to be satisfied that a statement was made voluntarily before they can rely on it (para 14):

“The reasoning of the members of the Appellate Committee was not identical on all points, but a majority held that in the light of section 76(2) of PACE a confession which was not made voluntarily could not be admitted in evidence, and a differently constituted majority held that the direction was inconsistent with the requirements of article 6(1) of the European Convention for the Protection of Human Rights and Fundamental Freedoms. St Vincent and the Grenadines has a statutory provision equivalent to PACE and their Lordships accordingly consider that in appropriate cases a direction should be given along the Mushtaq lines, that the jury should not rely on a confession which they think has or may have been obtained by maltreatment or oppression, even if they conclude that its contents were true.”

This common law power of the jury to reject a statement as a result of their own assessment of its voluntariness, is in parallel with the judge’s duty to exclude the statement from evidence if he is not satisfied beyond reasonable doubt that it was made voluntarily. The jury here is, in effect, reviewing the judge’s decision as to admissibility. The need for the jury to be directed about this has not been addressed in the Evidence Act 2006[NZ], so, apparently, it remains as a common law requirement.

Lawfulness, arbitrariness and road blocks

When the police, in an effort to catch someone who has committed a serious crime, set up a road block and question motorists, are they subjecting the occupants of the cars to arbitrary detention if there is no statutory authorisation for it? This question was considered in R v Clayton [2007] SCC 32 (6 July 2007).

The Supreme Court of Canada unanimously restored the convictions of the two respondents for offences involving firearms. Within minutes of receiving a “911” call concerning apparent armed offending in a car park, police set up a roadblock at its exit. Although the respondents’ car did not match the description given by the caller, it was the first and only car to leave. Clayton, the passenger, was wearing gloves, although the weather was warm. Clayton and the driver, Farmer, were consistent with the description of the offenders that the caller had given, in that they were “black males”. Evasive behaviour by Clayton contributed to the police decision to search the car.

Although the result was unanimous, the Court split in its reasoning. The majority judgment, by Abella J, of herself and McLachlin CJ, Bastarache, Deschamps, Charron and Rothstein JJ reasoned that the detention in this case was lawful at common law so there was no violation of the Charter right against arbitrary detention.

The other judges, Binnie (who delivered their judgment), LeBel and Fish JJ held that, yes, the detention here was lawful at common law, but it was still necessary to consider arbitrariness under the Charter; since there was a breach in that the detention was arbitrary (because everyone was to be stopped), the next question was whether this was a justified limitation (s 1) on the right not to be arbitrarily detained. Application of Charter jurisprudence (R v Oakes [1986] 1 S.C.R. 103) indicated that this limitation was justified.

Both approaches involve balancing exercises. Abella J pointed out (para 21) that there is a consistency between common law powers and Charter values, because under each the state must justify interference with liberty. The common law balancing, pursuant to LeDain J’s dictum in Dedman v. The Queen, 1985 CanLII 41 (S.C.C.), [1985] 2 S.C.R. 2, 35 requires that

“The interference with liberty must be necessary for the carrying out of the particular police duty and it must be reasonable, having regard to the nature of the liberty interfered with and the importance of the public purpose served by the interference.”

The Court in Clayton unanimously held that under this test the police action here was lawful at common law (Abella J at 32, Binnie J at 94).

At that point, Abella J held that there was therefore no arbitrary detention (para 41). This reflected her initial proposition (para 19) “If the police conduct in detaining and searching Clayton and Farmer amounted to a lawful exercise of their common law powers, there was no violation of their Charter rights.”

Binnie J, however, thought differently. His position (para 101) can be summarised as, some detentions may be lawful because they are arbitrary but justified limitations under s 1 of the Charter, so just because a detention is lawful does not mean that it is not arbitrary. Here is what he said:

“101   On the facts, I believe the police roadblock in this case, although authorized by the common law, did infringe the respondents’ rights under s. 9 of the Charter, which states as follows:

          9. Everyone has the right not to be arbitrarily detained or imprisoned.

An arbitrary detention can be upheld if found to be prescribed by a law which in itself constitutes a reasonable limit under s. 1 (see Hufsky [1988] 1 S.C.R. 621 and Ladouceur [1990] 1 S.C.R. 1257). It is in this sense, I believe, that the obiter in Mann [2004] SCC 52 that “a lawful detention is not ‘arbitrary’” (para. 20) should be understood. A detention, though arbitrary, may still be constitutional if the law on which the detention rests is a “reasonable limit” within the meaning of s. 1. The specific point in Mann itself was that a detention based on individualized suspicion is based on rational criteria and is not, therefore, arbitrary.”

This reasoning assumes that the notion of arbitrariness has not already been taken into account in the common law balancing exercise. What amounted to arbitrariness here, as Binnie J put it (para 103) was “…there was no individualized suspicion of them or the vehicle in which they were travelling (which did not match the 911 caller’s description) or other criteria to “tailor” the roadblock more precisely.”

One would have thought that the Dedman test, above, encompasses those matters. Binnie J did address this, however, saying (para 78):

“I do not believe the Dedman analysis and the s. 1 analysis are duplicative of one another. While both involve a measure of balancing, Dedman is a creature of the common law, and nothing is said explicitly in that case about the presumptive paramountcy of “the liberty interfered with” or putting the onus on the Crown to “demonstrably justify” the measure as a “reasonable limit”. Nor does Le Dain J. speak explicitly of such concepts as minimal impairment. The Charter standard is higher. The pre-Charter common law position was more loosely framed in recognition perhaps of the reality that relevant evidence would generally have been admissible even if the police conduct was unauthorized: R. v. Wray, 1970 CanLII 2 (S.C.C.), [1971] S.C.R. 272; Hogan v. The Queen, [1975] 2 S.C.R. 574.”

That is to take a rather static view of the common law balancing exercise, treating it as something to be done as if the Charter did not exist. Binnie J’s analysis produces the potentially confusing concept of the “lawful but arbitrary” detention. We might wonder, also, whether the detention here was really “arbitrary”: to detain for brief questioning all the people leaving the scene of a crime in the circumstances of this case did not involve the capriciousness or randomness that “arbitrary” connotes. Having found the search lawful at common law because it was reasonable, the conclusion should have been that there were no Charter issues. That was Abella J’s reasoning.

Of course Binnie J concluded that the detention was, although arbitrary, justified under s 1 of the Charter. There is merit in considering all rights balancing within the framework of the Charter, where the balancing process has received greater analysis than at common law. While this case is relatively uncontroversial, because of the potential public safety concerns, where the facts are innocuous the balancing can be controversial, and the influence of a Bill of Rights may not make the balancing exercise easier, as we saw in discussing Brooker v R [2007] NZSC 30 (blogged here, 4 May 2007).

Eliminating floppiness?

Vagueness in the criteria for the admission of evidence may allow a judge to take account of the accused’s right to a fair trial. This sort of justification for vagueness was advanced in Ahern v R (1988) 165 CLR 87, at para 17:

“The aim in limiting the use which might be made of a co-conspirator’s acts or declarations is to exclude such evidence when its admission might operate unfairly against an accused. For this purpose, the element of discretion implicit in the terms ‘reasonable evidence’ is desirable.”


That threshold of reasonable evidence refers to the non-hearsay evidence of the existence of a common intention and of the accused’s participation in the carrying out of that intention, in one of the more complicated of the common law exceptions to the rule against hearsay.

This exception to the hearsay rule (known variously as the co-conspirators exception, the preconcert exception, or the common enterprise exception) has a number of components. The hearsay component is a statement made by a person, not available as a witness, about what the accused would be doing in furtherance of the relevant common purpose. If the other requirements of this exception are met, this is the statement that becomes admissible evidence. As with all hearsay evidence, the law imposes a requirement of reliability on such a statement before it can be admissible. In addition, the co-conspirators rule has a precondition concerning the non-hearsay evidence in the case, which also applies before the hearsay statement can be admitted in evidence. This is, proof (to a degree which was the subject of today's decision in Jiang, discussed below) that there was a common purpose and that the accused was a party to it. A further limitation imposed by the co-conspirators rule is that only hearsay evidence which demonstrates the accused’s participation in the furtherance of the common purpose is admissible; narratives of what the accused did in the past do not usually satisfy this “in furtherance” requirement.

There has been a difference of judicial opinion over whether the threshold applicable to the non-hearsay evidence in this exception should be reasonable evidence, or proof on the balance of probabilities. The difference has usually been acknowledged to be of little practical significance in the cases where this point has arisen, but, nonetheless, some judges have supported the balance of probabilities formulation. Indeed, until today, the threshold in New Zealand was the balance of probabilities. Canada uses the balance of probabilities: R v Carter (1982) DLR (3d) 385 (SCC) and R v Mapara [2005] SCC 23 (blogged here 4 May 2005).

Today, the Supreme Court in Jiang v R [2007] NZSC 51 (also called Qiu v R) held that the threshold applicable to the co-conspirators’ exception is reasonable evidence. The Court noted that this criterion is in "harmony" with the provisions concerning the admissibility in the Evidence Act 2006 (enacted, but not yet in effect). Section 18(1)(a) of that Act requires, as one of the conditions of the general exception to the exclusion of hearsay statements, that “the circumstances relating to the statement provide reasonable assurance that the statement is reliable”. This, of course, is a reference to the reliability of the hearsay statement.

This “harmony” point is an agreeable consistency, but it is not compelling, because harmony between the reliability of the hearsay statement and the non-hearsay threshold is not strictly necessary. They are quite distinct requirements.

Under the original form of the Act (see below for reference to new s 12A) it would have been necessary to ask what is “reasonable assurance” of reliability of the hearsay statement. That could be indicated by non-hearsay proof on the balance of probabilities of the existence of the common intention, plus the other requirement of the common law exception, namely that the statement was made in furtherance of the common design. Or it could, as the Court held, be indicated by reasonable non-hearsay evidence of the common purpose. I use the word “indicated” here, because it is possible that a hearsay statement concerning the accused’s participation in the joint enterprise could carry reasonable assurance of reliability for reasons other than that the non-hearsay evidence reaches a threshold. It may be that, if that is the position, then the co-conspirators exception does not need to be relied on to achieve admission of the hearsay statement.

In any event, the Evidence Act 2006 is to be changed from the way it was when the Supreme Court decided Jiang. A new section, 12A is to be inserted, providing that the common law rules concerning co-conspirators statements are not to be affected by anything in the Act. Of course, one view is that the common law may be evolved to bring it into line with the general hearsay exception in the Act, but it is also possible to infer a legislative expression of confidence in the common law rules in their form as at the date of this (now, future) amendment.

Another point which is unclear is whether reasonable evidence means the same thing all the time, or whether it is variable, sometimes being more than proof on the balance of probabilities and sometimes less.

In argument in Jiang (SC41/2006, 15 February 2007) Tipping J raised the question whether reasonable evidence is more or less than the balance of probabilities (p 32 of the transcript):

“… [the trial judge] then talks about balance of probabilities as if that was the higher standard. I think she takes the view that reasonable evidence is a lower threshold than balance of probabilities which is a debatable point.”

Blanchard J added (p 33):

“… I wasn’t sure which is higher than the other. I actually think I prefer reasonable evidence which is safe to admit because it’s a more direct expression. It tells you what to look at whereas balance of probabilities is a bit floppy in this context.”

These concerns were not addressed in the Court’s judgment in Jiang. It seems that, where appropriate, the requirement of reasonable evidence will be higher than the balance of probabilities.

Unappealing confusion

The High Court of Australia’s attempt in Weiss v R (2005) 224 CLR 300 (blogged here 16 January 2006) to clarify the law concerning the application of the proviso has not prevented differences of opinion as to the results of appeals. The same Court, indeed, a Bench of judges who had all participated in the Court’s judgment in Weiss, has divided 3 – 2 on whether to apply the proviso in Libke v R [2007] HCA 30 (20 June 2007). The majority decision in Libke was to dismiss the appeal.

Weiss held that the appellate court is to apply the words of the legislation that enacts the proviso, and is not to introduce considerations of what a jury would do in the absence of the error that had occurred at trial. I suggested in January 2006 that this means that “ … the appellate court acts as a substitute jury, albeit one that has not seen or heard the witnesses.”

In Libke the point decided in Weiss was summarised by Hayne J (with whom Gleeson CJ and Heydon J concurred) at para 115 as:

“…the unanimous decision of this Court in Weiss v The Queen … warned against attempting to describe the operation of the statutory language in other words, lest such expressions mask the nature of the appellate court's task in considering the application of the proviso. The Court expressly discountenanced [para 35] any attempt to predict what a jury (whether the jury at trial, or some hypothetical future jury) would or might do. Rather, the Court said [para 35] that “in applying the proviso, the task is to decide whether a 'substantial miscarriage of justice has actually occurred'". Unless, and until, a majority of this Court qualifies what is said in Weiss, the intermediate courts of Australia must continue to apply that decision.”

Application of an authority at the same time as mentioning the possibility of it being overturned is hardly a sign of great confidence in its correctness. Indeed, it is virtually impossible for an appellate court to assess whether a substantial miscarriage of justice may have occurred without reference to what a jury might have decided in the absence of the error at trial. Hayne J does this himself (para 113):

“…the question for an appellate court is whether it was open [Hayne J’s emphasis] to the jury to be satisfied of guilt beyond reasonable doubt, which is to say whether the jury must [Hayne J’s emphasis], as distinct from might [Hayne J’s emphasis], have entertained a doubt about the appellant's guilt [footnote citing M v The Queen (1994) 181 CLR 487 at 492-493]. It is not sufficient to show that there was material which might have been taken by the jury to be sufficient to preclude satisfaction of guilt to the requisite standard. In the present case, the critical question for the jury was what assessment they made of the whole of the evidence that the complainant and the appellant gave that was relevant to the issue of consent to the digital penetration that had occurred in the park. That evidence did not require the conclusion that the jury should necessarily have entertained a doubt about the appellant's guilt.”

Aside from references to a jury (or, "the" jury), this looks like a burden on the appellant (and, of course, on appeal there is a burden on the appellant) to show not just that there was a miscarriage of justice but also to show that it was substantial so as to prevent the application of the proviso. The minority of the Court, Kirby and Callinan JJ, held that, once a miscarriage of justice had been demonstrated, the burden was on the respondent to show that it was not substantial (para 44, citing Mraz v The Queen (1955) 93 CLR 493 per Fullagar J at 514, a passage that received a bit of a knock in para 27 of Weiss).

The minority emphasised its rehabilitation of Fullagar J’s dictum in Mraz by citing Stokes v The Queen (1960) 105 CLR 279 at 284-285 per Dixon CJ, Fullagar and Kitto JJ, saying (para 45):

“As it is put in Stokes v The Queen …, an appellate court should only apply the proviso if the irregularity "could not reasonably be supposed to have influenced the result". If this cannot be ruled out, it may be impossible for a court to be satisfied that a substantial miscarriage of justice has not occurred. …”

Even while differing in its interpretation, the minority in Libke did not seek to modify Weiss. On occasions such as this, where the time is not quite right to modify an earlier decision, a minority can at least emphasise the salvageable bits (para 43 - 44):

“In Weiss, after reviewing the history of the statutory demise of the Exchequer rule in criminal cases and pointing out that in consequence, an appellate court was not obliged in all cases in which irregularities had occurred to uphold an appeal, the Court said this [para 45]:

"Likewise, no single universally applicable criterion can be formulated which identifies cases in which it would be proper for an appellate court not to dismiss the appeal, even though persuaded that the evidence properly admitted at trial proved, beyond reasonable doubt, the accused's guilt. What can be said, however, is that there may be cases where it would be proper to allow the appeal and order a new trial, even though the appellate court was persuaded to the requisite degree of the appellant's guilt. Cases where there has been a significant denial of procedural fairness at trial may provide examples of cases of that kind."


“44. The foregoing statement must be read with the several others [footnote citing: Driscoll v The Queen (1977) 137 CLR 517 at 524-525; Festa v The Queen (2001) 208 CLR 593 at 627 [110]; and TKWJ v The Queen (2002) 212 CLR 124 at 144-145 [68]] made in this Court which emphasise that, once it is shown, as it has been to all members of this Court, that irregularities disadvantageous to the appellant occurred at his trial, it is for the prosecution to satisfy the appellate court that such irregularities have caused no substantial miscarriage of justice. This is clear from the oft cited passage of Fullagar J in Mraz v The Queen …”


There are clear points of contrast here with the approach to the proviso in New Zealand: see, for example Bain v R (New Zealand) [2007] UKPC 33 (blogged here, 11 May 2007) at paras 34 – 38, referring to well settled principles which were not challenged on that appeal. The error in Bain was that the appellate court had substituted its judgment about the appropriate verdict for that of the jury. Perhaps most of the judges on the High Court of Australia would not have seen that as objectionable.

Probative value and prejudicial effect

The discretion to exclude evidence when its probative value is outweighed by its illegitimately prejudicial effect has given rise to some difficulties in predicting the admissibility of evidence. The discretion has been enacted in s 8 of the Evidence Act 2006 [NZ]. Here are my observations on how this potentially difficult section should be interpreted.

Domestic harmony and judicial disharmony

Where a witness is not competent to be a prosecution witness without the consent of the accused, can hearsay evidence from that person nevertheless be admissible if it is reliable? Both these common law rules received consideration in R v Couture [2007] SCC 28 (15 June 2007), and the Supreme Court of Canada split 5 – 4. The minority accused the majority of departing from the approach established in R. v. Hawkins, 1996 CanLII 154 (S.C.C.), [1996] 3 S.C.R. 1043, which was that the competency rule may be used to exclude hearsay evidence that would otherwise be admissible if it was unfair to admit the evidence. Interpretations of the judgments that were delivered in Hawkins differed in Couture. Essentially, the difference is between whether or not the exclusion of the evidence of an incompetent witness is a matter for residual discretion, or whether it is a consideration independent of the criteria for hearsay admission. The majority in Couture, in a judgment delivered by Charron J, approached the competence issue independently of the hearsay issue (para 63), whereas the minority (in a judgment delivered by Rothstein J) regarded it as a residual matter (para 105).

This interesting interplay of two common law rules arose at the accused’s trial for murder. The prosecution wished to adduce evidence from his spouse in the form of statements she had made, during a period of matrimonial discord, to the police, in which she said her husband had admitted the killings to her. Being the accused’s spouse, she was not competent as a prosecution witness under Canadian common law, without the accused’s consent. The prosecution argued that the statements were sufficiently reliable to be admitted pursuant to the “principled exception” to the hearsay rule (also known, in its New Zealand version, as the residual exception, although here the term “residual” is a bit confusing, so I will use the Canadian expression), and that there was no reason to exclude them as marital communications under the residual discretion to exclude evidence if its probative value would be outweighed by its illegitimately prejudicial effect or if it would be unfair to admit the evidence.

The majority examined the underlying rationales for the existence of the spousal incompetency rule (para 43):

“The first justification for the rule is that it promotes conjugal confidences and protects marital harmony. The second is that the rule prevents “the indignity of conscripting an accused’s spouse to participate in the accused’s own prosecution” (Hawkins, at para. 38). Wigmore describes this second justification as the “natural repugnance in every fair-minded person to compelling a wife or husband to be the means of the other’s condemnation, and to compelling the culprit to the humiliation of being condemned by the words of his intimate life partner” (Wigmore on Evidence (McNaughton rev. 1961), vol. 8, at §2228, p. 217 …).”

Noting the controversy over whether the rule should be changed, the majority confronted the role of the courts as compared with the legislature, quoting (para 47) the following from Iacobucci J in R. v. Salituro, 1991 CanLII 17 (S.C.C.), [1991] 3 S.C.R. 654:

“Judges can and should adapt the common law to reflect the changing social, moral and economic fabric of the country. Judges should not be quick to perpetuate rules whose social foundation has long since disappeared. Nonetheless, there are significant constraints on the power of the judiciary to change the law. As McLachlin J. indicated in Watkins, supra, in a constitutional democracy such as ours it is the legislature and not the courts which has the major responsibility for law reform; and for any changes to the law which may have complex ramifications, however necessary or desirable such changes may be, they should be left to the legislature. The judiciary should confine itself to those incremental changes which are necessary to keep the common law in step with the dynamic and evolving fabric of our society. [p. 670]”

In New Zealand, the legislature has abolished the spousal non-compellability rule (as it had existed here), making spouses both eligible and compellable as witnesses: Evidence Act 2006, s 71. The New Zealand Law Commission, proposing this change, had observed (NZLC R55, Vol 1, para 342 – 343) that non-compellability could not be supported by logic or policy, citing the same edition of Wigmore (at para 2228) as that cited by the Couture majority, above, for the proposition that marital immunity is a mere anachronism and an indefensible obstruction to truth.

For the majority in Couture, Charron J concluded, on the incompetence point (para 71):

“…If this Court were to rule that statements made by spouses can be admitted at trial based solely on threshold reliability without further regard to the spousal incompetency rule, I agree with Mr. Couture’s contention that this would encourage the institutionalized taking of spousal statements for the express purpose of introducing them at trial, a practice that would seriously undermine the preservation of marital harmony. This result would constitute a significant inroad on Parliament’s policy choice to maintain the rule against spousal incompetency, a result not intended by the majority in Hawkins. For that reason, I would conclude that this factor alone is sufficient to distinguish this case from Hawkins. The operation of the principled approach to the hearsay rule would effectively thwart the spousal competency rule and, consequently, cannot provide a basis for admitting the evidence in this case.”

This made it unnecessary to consider what the position would have been if the principled exception to the hearsay rule applied. The policy of avoiding risk to marital harmony was dominant. However, since the other justification for the incompetence rule, namely the indignity of having one spouse participate in the other’s prosecution, became apparent in the light of the hearsay exception, that was considered too. In the circumstances of this case, the defence could only challenge the reliability of the spouse’s evidence by calling her as a witness (she being competent and compellable for the defence), and that would raise this second justification for the incompetence rule.

Here the trial judge’s application of the principled exception to the hearsay rule was wrong, and reviewable on appeal, in two respects: the judge had regarded the wife’s repetition of the alleged confessions to other people as corroboration (whereas a person cannot corroborate himself), and the judge had reversed the onus, effectively requiring the defence to show why the hearsay was unreliable (paras 83 – 85). This allowed the Supreme Court to substitute its own assessment, and the majority held that there were, here, no adequate substitutes for the inability to cross-examine the witness in testing the reliability of the evidence (para 91). It was not the sort of evidence that even a sceptical person would regard as trustworthy (para 101).

The minority differed on this point, holding that the evidence was sufficiently reliable to be admissible, and the absence of cross-examination insufficiently important to require its exclusion. The residual discretion therefore, on the minority’s approach, came into play: was there unfairness sufficient to outweigh the probative value of the evidence? The only unfairness that could arise here would be from the stress that the giving of the evidence would place on the defendant’s marriage, and it was held (para 142) that was minimal.

This difference in approaches to the hearsay exception calls to mind my discussion of the admissibility of hearsay confessions (see paper linked in blog for 28 May 2007). I argued there that the reliability threshold for the admission of hearsay evidence was in danger of being eroded if reliance was placed instead on the discretion to exclude improperly prejudicial evidence. Couture, itself a hearsay confession case, saw a split in whether the evidence was sufficiently reliable to be admissible as hearsay. Such a difference is of itself an illustration of why the reliability requirement should be applied rigorously. The rule should be strong because of the potential variety in assessments of the facts.

Clarity and obscurity

One of the things that makes law both difficult and interesting is the shift between clarity and obscurity. Authoritative statements of the law, expressed clearly and simply, and in apparently absolute terms, tend later to be qualified by cases in which those requirements are not treated as absolute.

Readers of Daniel and others v The State of Trinidad and Tobago [2007] UKPC 39 (13 June 2007) will, no doubt, be reminded of this characteristic of the law. In addition to the clarity/obscurity shift, there can also be a failure to offer a precise statement of law notwithstanding that the opportunity for such has arisen. Here, the Board declined (para 34) to formulate rules subsidiary to the requirement that a person in custody be informed of his right to legal advice. One of the appellants, being illiterate and therefore probably at some disadvantage, had been told of his right to a lawyer but had not been told how he might exercise that right. The Board held that it was a matter for the judge in the circumstances of each case to decide whether the defendant had been properly informed of his right to legal advice. In contrast, the approach that the Board took to this right in Ramsarran v Attorney-General of Trinidad and Tobago [2005] UKPC 8 (blogged here 28 February 2005) emphasises the social utility in facilitating the right to legal advice. That case, not cited in Daniel, concerned arrest on warrant for outstanding fines, whereas this involved murder.

Another point of appeal raised in Daniel was the omission of a good character direction, caused by the defendant’s counsel’s failure to raise the matter. This is something, the effect of which depends on the circumstances of each case; there are some cases where omission of mention of the defendant’s good character can result in the quashing of a conviction: see the blog on Gilbert v R 29 March 2006. In Daniel the omission of the point was held not to have prejudiced the fairness of the trial, as the defendant had been in custody awaiting trial on other serious charges at the time of his trial for murder. The Board held that the risk of this being disclosed to the jury if character were put in issue made counsel’s omission understandable.

An interesting, although not novel, point about secondary liability was mentioned in Daniel. Departure from a joint enterprise (or common intention) by one offender, who embarks on the commission of an offence outside that joint enterprise, can be followed by a decision by another offender to assist him in the commission of that new offence. The Board said this about secondary liability (para 38):

“The concepts of joint enterprise and aiding and abetting cannot be placed into watertight compartments. A defendant may have joined an enterprise to commit one crime, only to find that his companions went beyond what he had contemplated and so in committing a different crime were acting outside the bounds of the joint enterprise. He may nevertheless have remained with them and lent assistance or encouragement to them in the commission of the new crime, which would make him a secondary party as an aider and abettor. The judge may have failed to appreciate this in categorising the appellants' participation exclusively as members of a joint enterprise in the case of Daniel and Archibald and as aiders and abettors in the case of Garcia and Marshall. This did not operate to their disadvantage, however, and if his directions were sufficient on the content of each concept and there was evidence on which they could properly so find, the jury were entitled to find the appellants guilty on the basis ascribed to each by the judge.”

Another topic raised in Daniel was inferences. The judge had not told the jury that they must exclude every inference consistent with innocence before they could find an accused guilty. (I will call that the exclusionary direction.) Instead, the normal direction on circumstantial evidence was given, amounting to little more than a reminder that guilt must be proved beyond reasonable doubt. The exclusionary aspect of the inferences direction was emphasised in Taylor v R (blogged 14 March 2006), and the need to consider inferences of innocence in the context of all the evidence in the case was stressed in R v Hillier (blogged 23 March 2007). These cases were not mentioned by the Board in Daniel. Whether the exclusionary direction needs to be given depends on what is appropriate in the circumstances of each case, as I noted in commenting on Taylor, but the vagueness of that approach means that the absence of an exclusionary direction on inferences will frequently become a matter for consideration on appeal.

The appeal of hypotheticals

On Wednesday Kirby J in the High Court of Australia made an interesting observation on the extent to which that Court should engage in speculation about the legal consequence of events that have not yet occurred:

“Avoiding hypothetical decisions: … experience teaches that it is ordinarily wise to withhold substantial comments on the consequences of future events for legal rights and duties. Those events (if they occur at all) have an unpleasant habit of following an unpredictable course, often quite different from that anticipated at an earlier time. From its earliest days to the present, this Court has been reluctant to proffer legal advice and predictions on the basis of hypothetical facts that have not yet arisen [footnote: The State of South Australia v The State of Victoria (1911) 12 CLR 667 at 674-675; In re Judiciary and Navigation Acts (1921) 29 CLR 257 at 267; North Ganalanja Aboriginal Corporation v Queensland (1996) 185 CLR 595 at 623, cf 666-668; Bass v Permanent Trustee Company Ltd (1999) 198 CLR 334 at 357 [49] per Gleeson CJ, Gaudron, McHugh, Gummow, Hayne and Callinan JJ]. In part, this reluctance has grown out of the constitutional requirement for there to be a "matter" before the Court. There is no problem in that regard in the present proceeding. The appellant's appeal is such a "matter". In part, it has arisen because of the Court's usual unwillingness to expand the ambit of the "matter" by offering predictions about the outcome of arguments of double jeopardy, in any future trial of the appellant, in advance of any such trial and without the concrete circumstances, enlivened by evidence, said to give rise to the peril of double jeopardy forbidden by the earlier judgment of acquittal. As recent divided decisions of this Court show, this is a tricky area of the law. It contains many difficulties [footnote: Pearce v The Queen (1998) 194 CLR 610; Island Maritime Ltd v Filipowski (2006) 226 CLR 328. See also R v Carroll (2002) 213 CLR 635].”

The case before the Court, AJS v R [2007] HCA 27 (13 June 2007) involved the power of a court of appeal to order a new trial. The point is not controversial, and essentially is that where a jury has been directed that, if it finds the accused guilty on one count, it should not consider whether or not the accused is guilty on an alternative count, and it does find him guilty of the first count, a court of appeal may, if it quashes the conviction and enters an acquittal on that count, order a new trial on the alternative count.

The fact that complex issues may arise, if the prosecution decides to proceed with the new trial on the remaining count, is irrelevant. Arguments about double jeopardy are immaterial. The other judges, in a joint judgment (making an order with which Kirby J agreed), did give some consideration to what might happen if such arguments were raised, and decided this point as follows:

“19. No question of double jeopardy arises in the present matter. The proceedings commenced by the prosecution against the appellant were, as the Court of Appeal's orders recognised, only partly determined by that Court's disposition of the appeal. The second of the offences now under consideration (the offence of committing an indecent act) was a statutory alternative to the first. There has been and would be no double prosecution of the kind considered in Pearce. In Pearce, the prosecution sought and obtained convictions for two offences charged in the one indictment. Further, unlike Island Maritime, there would be no separate institution of a second prosecution. In this case the prosecution does not seek to institute new and different proceedings against the appellant after the final determination (against the prosecution) of earlier proceedings. The charge of incest preferred against the appellant has now been finally resolved in his favour. He is entitled to the entry of judgment and verdict of acquittal of that offence. But the other, lesser, statutory alternative offence of committing an indecent act put in issue by the presentment charging the appellant with incest has not been determined by the Court of Appeal and remains unresolved.”

Whether that is an acceptable approach to what constitutes double jeopardy is arguable. It seems beside the point to rely on procedural niceties, saying the proceedings hadn’t been determined. At the trial there was a fundamental error (leaving to the jury a charge on which there was insufficient evidence) which caused the jury to fail to consider the alternative charge. From the accused’s point of view, he was in jeopardy of conviction on either charge as soon as the trial started, and this jeopardy continued until the jury returned its verdict. The prosecution did not seek to withdraw either charge, and would have been happy with a conviction on one or the other. It was not the accused’s fault that the trial was completed unsatisfactorily. The double jeopardy rules (available to the accused in the form of the special pleas of autrefois convict or autrefois acquit, or as an application to stay the second proceedings) stem from the need to prevent an abuse of the court’s process. Such an abuse would, arguably, occur if the accused had to undergo a new trial because of this sort of systemic failure.

On the other hand, errors often occur which, when they are corrected on appeal by the quashing of convictions, lead to a new trial being ordered. Such new trials are not counted as double jeopardy. Often this occurs because the judge makes a mistake when summing up the case to the jury. These sorts of errors, certainly not the fault of the accused, result in the accused having to undergo another trial. At that trial the prosecution may be able to present a stronger case as a result of having had weaknesses highlighted by the accused at the first trial. That is not usually thought of as double jeopardy. Plainly, the policy underlying the double jeopardy doctrine needs to be clearly articulated and its applications rationalised.

Kirby J’s reservations about deciding the double jeopardy point, which is hypothetical on the present case, partly focused on the fact that the High Court has reserved its decision in a case in which the meaning of giving the accused the “full benefit” of an acquittal has to be decided (Washer v The Queen, reserved 27 April 2007). His point appears to be that it would be unfortunate if the reasoning in the joint judgment in the present case (where full argument on particular facts had not been heard) were to be used as a precedent for deciding the case that is currently reserved.

Bias and sensitivity

It seems that a reasonable apprehension that the judge is biased will be sufficient grounds for a rehearing, regardless of whether the judge’s decision was correct. This is a reflection of the absolute nature of the accused’s right to a fair hearing: a conviction resulting from an unfair hearing will be quashed. There are many examples of cases where this has occurred, without the appellate court investigating the merits of the conviction, as the index to these blogs reveals. I discussed the High Court of Australia’s decision in Antoun v R on 8 February 2006, and suggested that

“We might, by now, anticipate that a case of apprehended bias would be a substantial miscarriage of justice, given the fundamental importance of the appearance of justice, but one must acknowledge that there is, at least, an argument that whereas actual bias would be a substantial miscarriage of justice, apprehended bias falls short of that in cases where there was no loss of a real chance of acquittal.”

As yet there is no example of a court making this, at least theoretically possible, distinction between apprehended and actual bias. A recent example of apprehended bias is R v Teskey [2007] SCC (7 June 2007). Here the Supreme Court of Canada was concerned with a case where a judge, sitting alone, had announced his verdict some four months after hearing the evidence, without giving reasons, and 11 months passed before the written reasons were given, by which time the defendant had appealed against his conviction.

The decision of the majority was delivered by Charron J. Apparent bias and actual bias were seen as equally objectionable:

“21 As reiterated in R v S. (R.D.) [1997] SCC 324, [1997] 3 SCR 484, fairness and impartiality must not only be subjectively present but must also be objectively demonstrated to the informed and reasonable observer. Even though there is a presumption that judges will carry out the duties they have sworn to uphold, the presumption can be displaced. The onus is therefore on the appellant to present cogent evidence showing that, in all the circumstances, a reasonable person would apprehend that the reasons constitute an after-the-fact justification of the verdict rather than an articulation of the reasoning that led to it.”

In Teskey, the facts were relatively complex and the case against the defendant was circumstantial. Delay in giving reasons would not, of itself, have been sufficient to sustain a claim of apparent bias, but other matters combined to support a reasonable apprehension of bias. These were (para 23):

•  the trial judge’s obvious difficulty in arriving at a verdict in the months following the completion of the evidence;
  
•  the absolutely bare declaration of guilt without any indication of the underlying reasoning;
  
•  the trial judge’s expressed willingness to reconsider the verdicts immediately after their announcement;
  
•  the nature of the evidence that called for a detailed consideration and analysis before any verdict could be reached;
  
•  the failure of the trial judge to respond to repeated requests from counsel to give reasons;
  
•  the contents of the reasons referring to events long after the announcement of the verdict suggesting that they were crafted post-decision;
  
•  the inordinate delay in delivering the reasons coupled with the absence of any indication that his reasons were ready at any time during the 11 months that followed or that the trial judge had purposely deferred their issuance pending disposition of the dangerous offender application.
  

In this appeal the Court of Appeal had had regard to the judge’s written reasons in upholding the conviction, and the Supreme Court held that the Court of Appeal should not have done that. Instead of remitting the case to the Court of Appeal for it to consider the question of the conviction on the basis of the record of the hearing, a retrial was ordered. Although the only indication of bias appears to have arisen from the post-hearing events, the judge-alone procedure carries with it a right to have reasons given for the decision, and those become part of the material proper for consideration on appeal. This was a case where it would seem to a reasonable person that the reasons were a justification for the verdict, rather than the verdict being a result of the reasons, and this approach would naturally have obscured any errors of reasoning that had led to the verdict. The appellant was therefore deprived of the opportunity to challenge the reasons for the verdict.

Abella J delivered the dissenting judgment of herself, Bastarache and Deschamps JJ. This placed emphasis on the presumption of judicial integrity and the high threshold required to displace that presumption. There was nothing in the case, in the opinion of the minority, to displace that presumption.

Is talk of presumptions and thresholds really appropriate in this context? Why should there be obstacles to exposing bias? Certainly, cogent evidence of bias should be required, but it does not follow that a high standard of proof of bias should be reached before the appellate court will look at the issue. Some people would see the conferment, by the judges at common law, of this sort of immunity on their judicial colleagues as rather sanctimonious; a bias against bias.

Looking overseas

Do rights-based restrictions on acceptable evidence collecting procedures apply to officials who act outside the jurisdiction of the country in which the trial is ultimately held? In R v Hape [2007] SCC (7 June 2007), a case that will be a leading authority on extra-territoriality and international criminal law, the Supreme Court of Canada decided that, while the Charter of Rights and Freedoms does not apply outside Canada’s territory, when the trial is subsequently governed by Canadian law questions regarding the admissibility of evidence are answered by reference to the acceptability of what happened judged in the light of the Charter.

By this approach, the fact that a course of conduct in the collection of evidence was lawful in the foreign jurisdiction would not be determinative of the admissibility of the evidence in a Canadian court. The ultimate criterion was, in this case, referred to as the “fair trial” requirement, although this usage of that expression must be read in context. The reference here is not to trial fairness in the technical sense of a trial in which the law is properly applied to facts that are determined without bias, but rather it is to fairness in the public policy sense: would admission of the evidence bring the administration of justice into disrepute?

Hape contains dicta on the interpretation of Charter provisions in view of Canada’s obligations in respect of international human rights law, particularly:

“56. In interpreting the scope of application of the Charter, the courts should seek to ensure compliance with Canada’s binding obligations under international law where the express words are capable of supporting such a construction.”

In this case, RCMP officers conducted searches of the defendant’s business documents held in the Turks and Caicos Islands, under supervision of an official of that country, and believing that they were acting lawfully. There was no evidence that what was done was not lawful, and in the circumstances the searches were not unreasonable in terms of s 8 of the Charter.

This approach is what was anticipated here, in noting the European Court’s decision re an Application by Saddam Hussein (see 17 March 2006). Questions still exist concerning what approach is appropriate to the possible use of torture in a foreign jurisdiction, resulting in the obtaining of evidence for use in a domestic court. This was considered here in commenting on the House of Lords case A v Secretary of State for the Home Department (see 9 December 2005). Neither of those cases was mentioned in Hape.

Secret appeals

Public interest may require that some information not be disclosed by the prosecution even though it might be of use to the defence in preparation for trial. Obviously, this can give rise to questions about whether the trial was fair. Further questions may arise about where these issues of fairness should be determined – by the trial judge, or on appeal?

In Botmeh and Alami v United Kingdom [2007] ECHR 456 (7 June 2007) the Chamber had to determine whether the Court of Appeal of England and Wales had correctly decided that failure to disclose information to the defence had not resulted in an unfair trial. The point, it should be noted, was not whether the information should have been disclosed, as that was a matter for the national court; the European Court was concerned only with whether the proceedings had involved a breach of Article 6 of the Convention.

At trial, the accused Palestinians had been charged with conspiring to make and detonate explosives in London. Some information had been disclosed to the defence prior to the trial, to the effect that the bombs may have been planted by an Iranian organisation. It seems that little was made of this at trial, although one appellant, in giving evidence-in-chief, had referred, for the first time, to another Palestinian who had some involvement in what may have been relevant events. Other than that oblique point, no attempt was made by the defence to make use of the information they had been given concerning the possible involvement of any specific person or organisation. The trial judge had, however, mentioned to the jury that they had to consider the possibility that persons other than the accuseds were involved.

After the accuseds were convicted and sentenced, the prosecution discovered that it had failed to reveal to the judge that it had more information about the possible involvement of a Palestinian group. Significantly, however, this information was to the effect that such involvement was discounted by the security services. The appeal to the Court of Appeal was on the grounds that the defence might have been assisted by knowing why that involvement had been discounted, because if the reasons were bad then that involvement might have been a live issue.

The Court of Appeal was therefore faced having to decide how to deal with the public interest immunity (PII) question: should it hear the matter (remitting it to the trial judge was difficult, as a long time had passed since the conclusion of the trial), and, if so, how should it proceed? Defence counsel were invited to see the confidential material if they would undertake to the court that they would not disclose it to the appellants. They took advice on this point, and concluded that they could not give such an undertaking. Instead of appointing a special advocate to see the material and make submissions on it independently of the appellants’ counsel, the Court of Appeal decided to examine the material and to hear ex parte submissions from the Crown. Then, the Court released a summary of the confidential material to defence counsel, and (para 43)

“ … The applicants were given a full opportunity to make submissions on the material which had been disclosed in summary form and on its significance to the issues raised by the case. On the basis of the submissions made, the Court of Appeal concluded that no injustice had been done to the applicants by not having access to the undisclosed matter at trial, since the matter added nothing of significance to what was disclosed at trial and since no attempt had been made by the defence at trial to exploit, by adducing it in any form before the jury, the similar material which had been disclosed at trial.”

The Chamber held that, in this case, the procedure adopted by the Court of Appeal had been correct, and that this remedied the failure to place the material before the trial judge.

In cases where, as here, the defence is that it was someone else who committed the offence, all the accused is expected to do is raise a reasonable doubt that that, indeed, might be so. Where the accused says he does not know who the offender was, it may be difficult for him to point to someone else as the likely suspect. Where the authorities have suspected someone else, but they later decide that, no, that other person was not involved, the accused could at least get from that a pointer about to whom to look for a reasonable doubt. An investigator might look into it for the defence, and find reason to suspect the other person. Here, however, the role of investigating this lead was undertaken by the security services, and they concluded that it was an unwarranted suspicion.

Are the security services an adequate substitute for an investigator selected by the defence? In the circumstances of this case, it was probably safe to assume that the security services would have no reason not to pursue anyone whom it considered there were good grounds to suspect. Once the security services could be trusted with this impartiality, the only issue would be their thoroughness in checking the reliability of their sources of information. That is a matter the Court could assess, without the need for full disclosure of the confidential information to the defence. The security services’ conclusion that someone else was not a likely suspect was something that the defence would be unlikely to be able to challenge. A reasonable doubt must, after all, be grounded in reality, and not be merely fanciful.

Hearsay confessions

A particularly tricky problem in evidence law is how to deal with the admissibility of hearsay confessions. These occur when a witness for the defence wishes to say that someone else, not available to give evidence, confessed to the crime with which the accused is charged. Here is a draft paper discussing a New Zealand Court of Appeal decision on this topic.

Hidden badness

An attack by an accused on the character of a prosecution witness, especially a complainant, may result in the court being informed of the accused’s bad character. In DS v HM Advocate [2007] UKPC D1 (22 May 2007) the accused challenged the validity of Scottish legislation to this effect, on the basis that it infringed his right to a fair trial.

Lord Hope noted that the accused’s right to a fair trial is absolute:

“17. … The Convention right that the appellant invokes is his right to a fair trial. This was described in Salabiaku v France (1988) 13 EHRR 379, para 28, as a fundamental principle of law. In Doorson v The Netherlands (1996) 22 EHRR 330, 358, para 67 the Court said that its task was to ascertain whether the proceedings as a whole were fair. In Dyer v Watson, 2002 SC (PC) 89, 113, I drew attention to the fact that a distinction must be made between those rights which are said by the Convention to be absolute and unqualified and those which are expressly qualified by provisions which permit them to be interfered with in certain circumstances. I said that the overriding right guaranteed by article 6(1) was a fundamental right which did not admit of any balancing exercise, and that the public interest could never be invoked to deny that right to anybody in any circumstances: see also Montgomery v HM Advocate, 2001 SC (PC) 1, pp 27E, 29F-G; Brown v Stott, 2001 SC (PC) 43, pp 60B 74B [also reported as Procurator Fiscal v Brown (Scotland) [2000] UKPC D3]. The fundamental nature of the right to a fair trial has been stressed repeatedly in subsequent cases both in the Judicial Committee and in the House of Lords: R v Forbes [2001] 1 AC 473, para 24; Porter v Magill [2002] 2 AC 357, para 87; Millar v Dickson, 2002 SC (PC) 30, para 52; Mills v HM Advocate, 2003 SC (PC) 1, para 12; Sinclair v HM Advocate, 2005 SC (PC) 28, para 37. The law-making powers of the Scottish Parliament do not permit it to pass laws which will deny an accused a fair trial.”

Under the relevant legislation, in the circumstances that arose in this case, it was for the accused to show that “the interests of justice” favoured non-disclosure of his prior conviction. As to this phrase, Lord Hope said:

“49. … the words "the interests of justice" should be read, in this context, as directed primarily to the accused's right to a fair trial. This issue should be addressed in the light of what I have already said about the reasons why previous convictions for sexual offences or an offence in which a substantial sexual element was present may be relevant. The objection should be tested in the light of what use may properly be made of the conviction with regard to the accused's propensity to commit the offence charged, and what use may properly be made of it with regard to his credibility if he were to give evidence or has made exculpatory statements before trial. The test needs to be exacting in proceedings on indictment, in view of the risk that the jury may attach a significance to the conviction which, due to its age or other factors, it cannot properly bear.”

Whether juries are likely to treat the accused’s previous convictions correctly is a matter that we have seen commented on in R v Becouarn [2005] UKHL 55 (blogged 5 August 2005), not cited in the present case, where a jury study was quoted. However, in DS the Privy Council took a robust approach to this point. Lord Rodger, with whom all the other members of the Board agreed, said:

“85 … [the section allowing the accused’s conviction to be revealed, once he had attacked the complainant’s character] would provide an element of parity or balance in the treatment of the two sides by giving the jury an opportunity, when considering their verdict, to have regard also to what the accused had done on other occasions. The balance between the two sides is not perfect: it is tilted in the accused's favour since the jury only get to know about his previous criminal sexual misbehaviour. Any other behaviour or any other aspects of his character or any condition or predisposition are not revealed.”

Baroness Hale also upheld the fairness of the legislative scheme:

“94 There is nothing intrinsically unfair in a court hearing evidence of an accused person's character and conduct, provided that it is relevant to something which the court has to decide. Our historic reluctance to trust the jury with this information arises from the fear that they may give it more weight than it deserves or regard it as proving that which it does not prove. The answer to that does not have to be to withhold it from them; they can be given clear and careful directions about how to use it.”

Lord Brown summarised the accused’s argument and answered it as follows:

“102 … having won the initial ruling that the evidence [of the complainant’s bad character] is required to enable him to defend himself properly, he submits that no inhibition should thereafter be put in his path; he should not be subjected to the pressure of having to choose between two evils: either forgoing the opportunity to advance his defence properly or allowing the jury to learn of his previous convictions of which otherwise they would have remained in ignorance.

“103 Plausible and beguiling though at first blush this argument may appear, it is to my mind founded upon a central fallacy. The long and the short of it is that the accused has no fundamental right to keep his past convictions from the jury. There is nothing intrinsically unfair or inappropriate in putting these into evidence and, indeed, in doing so not merely on the limited basis that they go only to the accused's credibility (the fiction which to my mind disfigured the administration of criminal justice in England and Wales for far too long, now at last ended by the Criminal Justice Act 2003—see particularly sections 101(1)(d) and 103(1)(a)) but on the wider ground that they bear also on the accused's propensity to commit offences of the kind with which he is charged.”

Great reliance is placed on the ability of the trial judge correctly to warn the jury about the use to which they may put the evidence of the accused’s bad prior conduct, and great reliance is also placed on the ability of juries to follow such directions. This latter reliance appears to be misplaced in the light of the jury study cited in Becouarn, above. Should decisions about the fairness of trials rest on legal assumptions that may not be justified in fact?

Howse of discontent

Another case (in addition to Bain, also decided on 10 May 2007, see blog below) in which a Court of Appeal thought that the case against the accused was strong enough to make errors at trial insufficient to amount to a substantial miscarriage of justice, is Bernard v The State (Trinidad and Tobago) [2007] UKPC 34 (10 May 2007).

Here, the court had compelled the accused to be represented at his murder trial by an inexperienced lawyer, who had only been admitted to the bar three months previously. There had been a failure to disclose material facts to the defence, as well as an understandable failure to make the most of the available forensic techniques that a more experienced counsel (“an older hand” para 25) would have had at his disposal.

The Board cited Randall v The Queen [2002] UKPC 19, [2002] 1 WLR 2237, 2251 as authority for the approach to take on the question whether errors at trial have resulted in unfairness, quoting para 28 of Randall:

“While reference has been made above to some of the rules which should be observed in a well-conducted trial to safeguard the fairness of the proceedings, it is not every departure from good practice which renders a trial unfair. Inevitably, in the course of a long trial, things are done or said which should not be done or said. Most occurrences of that kind do not undermine the integrity of the trial, particularly if they are isolated and particularly if, where appropriate, they are the subject of a clear judicial direction. It would emasculate the trial process, and undermine public confidence in the administration of criminal justice, if a standard of perfection were imposed that was incapable of attainment in practice. But the right of a criminal defendant to a fair trial is absolute. There will come a point when the departure from good practice is so gross, or so persistent, or so prejudicial, or so irremediable that an appellate court will have no choice but to condemn a trial as unfair and quash a conviction as unsafe, however strong the grounds for believing the defendant to be guilty. The right to a fair trial is one to be enjoyed by the guilty as well as the innocent, for a defendant is presumed to be innocent until proved to be otherwise in a fairly conducted trial.”

And added (para 29 of Bernard):

“There are statements in the Australian case of Wilde v The Queen (1988) 164 CLR 365 which, if taken out of context, could give support to a proposition that where the evidence against a defendant is overwhelmingly strong, the defects in procedure required for setting the verdict aside on the ground that the trial was unfair have to be such that there has scarcely been a trial at all. The Board applied the decision in Wilde v The Queen in the New Zealand appeal of Howse v The Queen [2005] UKPC 30, but it is not to be taken to have approved this formulation as the universally necessary criterion for proof of unfairness of a trial. In the context of the incorrect admission of evidence, the strength of the rest of the evidence will be material, but in a case of procedural unfairness their Lordships would regard the statement which they have quoted from Randall v The Queen as the appropriate approach. Determination of such an issue involves weighing the seriousness of the irregularities. If the defects were relatively minor, the trial may still be regarded as fair. Conversely, if they were sufficiently serious it cannot be accepted as fair, no matter how strong the evidence of guilt. In such a case it may also be said that the defendant was deprived of his constitutional right of due process.”

The distinction made here is between cases, like Howse, where the error at trial had been the wrongful admission of evidence, and cases like the present appeal, where the question was whether the error had involved procedural unfairness.

Whether this is a clear distinction remains to be seen. The wrongful admission of evidence could result in the fact-decider undertaking a biased task, so creating procedural unfairness, just as failure to disclose material facts to the defence, or failure to cast the doubt on the prosecution case that could have been cast, also gives rise to procedural unfairness.

Perhaps the Privy Council is, in Bernard, tactfully disagreeing with the contentious decision in Howse, where the Board was split 3-2. This, however, may be doubted, because this judgment was delivered by Lord Carswell, who had delivered the majority judgment in Howse. However, a reader of these decisions who applies the rule-of-thumb “Lord Bingham is never wrong” will note his absence from Howse and his presence in Bernard, in which only one judgment was delivered. Did Lord Bingham prompt this subtle qualification of the application of Wilde?

Justices Marple, Holmes and Poirot

Stating the law is one thing, applying it another. An appellate court may be able correctly to state the way it should approach the question of whether there has been a substantial miscarriage of justice in the case before it, yet, when it comes to deciding the appeal, it may contravene its own statement of the law.

This fundamental error occurred in the New Zealand Court of Appeal’s decision which yesterday was overturned by the Privy Council: Bain v R (New Zealand) [2007] UKPC (10 May 2007).

The law as to how the question of substantial miscarriage of justice should be determined was set out by the Court, and quoted by the Board (para 35), as follows:

"The court went on, in paragraph 24 of its judgment, to observe that when fresh evidence is admitted, it must move on to the next stage of the enquiry
“which is whether its existence demonstrates there has been a miscarriage of justice in the sense of there being a real risk that a miscarriage of justice has occurred on account of the new evidence not being before the jury which convicted the appellant. Such a real risk will exist if, as it is put in the cases, the new evidence, when considered alongside the evidence given at the trial, might reasonably have led the jury to return a verdict of not guilty.”

“The court pointed out (paragraph 25) that its concern is whether the jury, not the court, would nevertheless have convicted had the posited miscarriage of justice not occurred. This was consistent with
“the fundamental point that the ultimate issue whether an accused person is guilty or not guilty is for a jury, not for Judges. The appellate court acts as a screen through which the further evidence must pass. It is not the ultimate arbiter of guilt, save in the practical sense that this is the effect of applying the proviso, or ruling that the new evidence could not reasonably have affected the result.”"

It was surprising, after such a precise statement of the appellate court’s role, that the Court of Appeal should then have embarked on a detailed analysis of the evidence and come to its own determination of the appellant’s guilt. As I observed in an article mentioning this case, “Proof, fairness and the proviso” [2006] NZLJ 156, 158 (copy available on this link): “ … we are left to wonder whether a jury would have reached the same conclusion at a new trial.”

The Privy Council held that only a jury could assess the impact of the fresh evidence that the defence had obtained. The Court of Appeal’s reasons for concluding that the appellant was guilty were based on assumptions and matters that had not been raised at trial. As to these (para 115):

“…The Board does not consider it necessary to review these points in detail, for three reasons. First, the issue of guilt is one for a properly informed and directed jury, not for an appellate court. Secondly, the issue is not whether there is or was evidence on which a jury could reasonably convict but whether there is or was evidence on which it might reasonably decline to do so. And, thirdly, a fair trial ordinarily requires that the jury hears the evidence it ought to hear before returning its verdict, and should not act on evidence which is, or may be, false or misleading. Even a guilty defendant is entitled to such a trial.”

And added (para 119):

“ … Where issues have not been fully and fairly considered by a trial jury, determination of guilt is not the task of appellate courts.”

The conclusion was that there had been a substantial miscarriage of justice (which became apparent only after the trial, when the fresh evidence came to light), that the convictions should be quashed and a retrial ordered.

The lesson here is that procedural fairness is of fundamental importance, and an improper focus on the apparent guilt of the appellant must not cause the appellate court to take over the role which is properly that of a jury.

As evidence of the temptation to overlook trial fairness and to focus on apparent guilt, we need only recall that in last year’s report by a retired High Court Judge on the likely occurrence of miscarriages of justice in trials in New Zealand, no criticism was made of the upholding of the Bain convictions.

Disorderly rights

Offences of disorderly behaviour, involving the simplest of facts, can provide the basis for interminable disputation among jurists. This affords students an introduction to the role of law, the need for certainty and the problem of definition of crime, the proper scope of judicial discretion, the nature of judgment, and the complicating influence of competing rights on the definition of the offence.

Also, such apparently simple cases can be occasions for judicial disagreement as they proceed up the court hierarchy. A dogged appellant can eventually succeed, as happened today in Brooker v R [2007] NZSC 30 (4 May 2007). In this case, the successful appellant represented himself before the Supreme Court, thereby illustrating what many may suspect: it’s the issues, not the lawyers, that matter.

The judges took 288 paragraphs to deliver their separate judgments, and the case was decided by a bare majority of 3 to 2. As Thomas J, one of the dissenters, pointed out (para 150), in all, 10 judges had considered the case, 7 of them were in favour of convicting, and only the 3 majority judges at this final appeal stage effected the quashing of the conviction. That, of course, is an illegal count, comparing – if one might carefully choose one’s fruit – apples with oranges.

For our purposes, the result of the case doesn’t matter. We are more interested in how judges differ in their approaches to deciding how the actus reus of the offence is affected by the existence of competing rights.

I should point out that vagueness in the definition of an offence is not unusual. Most offences can be attempted, and the attempt is a separate offence. To be guilty of an attempt, one must perform an act that is sufficiently proximate to the commission of the full offence to constitute an actus reus of an attempt. Proximity is vague, and fact-dependent. Nevertheless, the courts have worked out various ways of asking whether particular circumstances disclose sufficient proximity. Those cases, certainly not always free of controversy, do not involve the added complication of conflicting rights.

In Brooker, the complication of competing rights was approached in two different ways. In the first, one right was seen as a limitation on the other. The complainant’s right to privacy was seen as a limitation on the accused’s right to freedom of expression, and the question was whether this limitation was justified. If it was, the defendant was guilty. This approach is evident in the majority judgments of Blanchard J (para 69), Tipping J (91).

The second way of dealing with the complication of competing rights does not involve justifying the limitation to the defendant’s right to freedom of expression. Instead, the competing rights are put against each other: the right to freedom of expression as against the right to privacy, and the question is which the balance favours. The dissenters, McGrath and Thomas JJ take this approach (paras 136, 231 respectively).

The other judge, Elias CJ, held that the lower courts had applied the wrong test for what disorderly behaviour means, failing to require a serious disruption to public order, and allowed the appeal because of that error. She added that the defendant could not have been convicted on the correct approach.

The two approaches, summarised above, to dealing with the complication of competing rights in the context of a vaguely defined offence, deserve some reflection. They reveal a difference in the idea of the role of rights in society. For Tipping J, the question was what a reasonable citizen should be expected to bear (para 91): how much disorderly behaviour should a reasonable citizen be expected to bear in the interests of upholding the defendant’s right to freedom of expression? The focus is on the impact on the defendant’s right, because the question is basically what is the proper scope of the criminal law? This may seem a little odd, since it is the defendant who is initiating the conflict. One might have thought that the question should be put as, what limitation on his right to freedom of expression should the defendant be expected to bear (by incurring criminal liability), in the interests of upholding the complainant’s right to privacy?

In contrast to rights limitation, the other approach, elaborated carefully by Thomas J, involves rights balancing. Pragmatists will appreciate his reference to the test of the “reasonable person” (para 199), echoing that introduced by the famous pragmatist Oliver Wendell Holmes Jr. Thomas’s pragmatism has been noted in these blogs before. Some, too, will appreciate his reference to Barak’s “The Judge in a Democracy” (para 177).

Central to this pragmatist view of rights are the ideas that everyone is entitled to be treated with equal concern and respect (para 171), and that the right to dignity is central to all human rights. Primacy is not accorded, as a starting point, to the rights of the defendant when they are in opposition to the rights of another citizen; this playing field is level (232). It is doubtful that, at this point, Thomas J was intending to suggest that all rights will have equal weight in a balancing exercise.

Theory is one thing, putting it into practice another. We do not have to agree with Thomas J’s conclusion on the facts of the case. It is difficult to form a view of these, because each judge, in reporting the facts of the case, puts them in a light which tends to support his or her conclusion. Such is the way of humans.

Questions remain about which of the two approaches, rights limitation or rights balancing, is appropriate to the decision whether behaviour constitutes an actus reus. The four judges who considered this split 2-2. The balancing approach is familiar in the context of determining the admissibility of evidence obtained improperly, where the public policy discretion arises from the need to prevent abuse of process and the related need to avoid bringing the administration of justice into disrepute. The definition of offences, however, has traditionally been based on utilitarian grounds, whereby, as JS Mill put it, the state is justified in restraining the freedom of those within its jurisdiction only to the extent that such restraint is necessary to prevent harm.

Multiple intents

When does an agreement to do one specific unlawful thing extend to encompass the doing of another unlawful thing? And when does a lawful purpose prevent liability for an unlawful purpose?

Two people might, for example, agree to stop traffic in order to prevent an accident at the scene of a breakdown. They have two intentions: to stop traffic, and to prevent an accident. Do these two intentions exist without qualifying each other, so that the legal consequences of one are unaffected by the existence of the other?

In DPP v Hurnam (Mauritius) [2007] UKPC 24 (25 April 2007) there is a reminder of these sorts of questions, although the case itself is a straightforward one. A criminal lawyer (the respondent) had been convicted of conspiring to hinder police by fabricating an alibi for a client, Bholah, in order to mislead the police in their investigation of a robbery. The Supreme Court of Mauritius had allowed his appeal against conviction on the basis that the prosecution had to prove an intention to hinder the police, as distinct from an intention to fabricate an alibi. The Privy Council restored the conviction.

“[20] …The Supreme Court was in their Lordships' view wrong to have regard only to the ultimate object or intention of the parties. A conspiracy may have several objects (cf the discussion in Williams, Criminal Law, The General Part, para 217) and one of the objects may constitute a means of achieving the ultimate object. As Webster J said in Lewis v Cox [1985] QB 509, 517 in the context of obstructing a police constable in the execution of his duty:

“" … a court is not obliged … to assume that a defendant has only one intention and to find what that intention was, or even to assume that, if he has two intentions, it must find the predominant intention. If, for instance, a person runs into the road and holds up the traffic in order to prevent an accident, he clearly has two intentions: one is to hold up the traffic, and the other (which is the motive of that intention) is to prevent an accident. But motive is irrelevant to intention in the criminal law …"

“In this case the parties intended to assist Bholah to escape criminal liability, and they did so by means of fabricating an alibi, which would have the foreseeable and intended result of diverting the police from investigating his actions, so hindering their enquiry into the commission of the robbery. The immediate intention of their agreed course of conduct was to hinder the police in the accomplishment of their ultimate aim of exculpating Bholah. That in their Lordships' view was an agreement to hinder the police, sufficient to constitute the conspiracy with which the appellant was charged. The fact that the respondent and Bholah had an ultimate aim of exculpating the latter does not invalidate that conclusion.”

The Board mentioned (para 23) a couple of general points, the approach to findings of fact in lower courts, and the need to demonstrate a clear departure from the requirements of justice when appeals are as of right, as is the case in Mauritius. The Board declined to interfere with the sentence of 6 months’ imprisonment, holding that the delay (from 2000) had been occupied largely by the trial and the appeals process, and observing that the importance of maintaining the high standards of conduct at the criminal bar supported the sentence.

This case, while not deciding any new point of law, serves as a reminder that the concept of agreement, as an element of conspiracy, can include more than one objective.

Judging the gatekeeper

In Wizzard v R (Jamaica) [2007] UKPC 21 (29 March 2007) the Privy Council was reminded of the decision of the House of Lords in R v Mushtaq [2005] UKHL 25, [2005] 1 WLR 1513 (blogged 22 April 2005). Although the facts of the present appeal did not require Mushtaq to be applied, some observations on it were made by the Board.

Because it is based on the ancient ("from the time of Lord Hale") common law principle against self-incrimination, Mushtaq was held to apply in Jamaica (and this implies, of course, that it extends to any other common law country unless there is legislation to the contrary effect), and to apply retrospectively to cases heard before Mushtaq was decided.

Mushtaq held that a jury must be directed, where the matter is in issue, that if they are not satisfied that a statement was made by the accused in the absence of oppression, they must ignore it.

Prior to Mushtaq, the Judicial Studies Board in England had approved a direction which, once the judge had decided the oppression point against the accused and ruled the statement admissible, the jury could give it what weight they wished, even if they thought it could have been made as a result of oppression. Thus, Mushtaq made an important correction to the way the law had developed.

Technically, what the jury is doing under the Mushtaq direction, is not deciding the admissibility question (which is a matter the judge had decided), but instead is deciding that, because of the possibility of oppression, the statement should be given no weight and it should be ignored. That, of course, is a highly technical distinction, but it has importance in preventing erosion of the standard that the judge must apply to the question of whether there was oppression.

It would be wrong for a judge to think that, because the jury can decide to ignore the statement, it may as well be ruled admissible even where there is a real risk that it had been obtained by oppression. The challenge for those involved with making judges apply the law with integrity is to ensure that this sort of error does not occur.

Inferences

An area of the law of evidence that has long been a subject of discussion and confusion is the drawing of inferences. In R v Hillier [2007] HCA 13 (22 March 2007) the High Court of Australia corrected an error of reasoning by an appellate court on this topic.

The accused had been convicted of murdering his estranged wife, his motive being alleged to be to obtain custody of their children. There was only circumstantial evidence of his guilt. The Court of Appeal (ACT) had quashed the conviction, holding ([2005] ACTCA 48 at [105]):

"other aspects of the evidence, such as that relating to the unusual features of the injuries she suffered and the apparent use of the handcuffs … [made] it difficult to reconstruct what actually occurred on the night in question and the evidence suggesting that another person may have been present at the time of her death".

The Court of Appeal found that this made it impossible to conclude, beyond reasonable doubt, that the accused was guilty.

As the High Court pointed out, it is wrong to isolate pieces of evidence and use them to support an inference inconsistent with guilt. Gummow, Hayne and Crennan JJ, jointly, with Gleeson CJ concurring, and Callinan J agreeing in allowing the Crown’s appeal (but who would have ordered a retrial, rather than, as the others did, remit the case to the Court of Appeal for reconsideration), held at para 46:

“The case against Mr Hillier was a circumstantial case. It has often been said that a jury cannot be satisfied beyond reasonable doubt on circumstantial evidence unless no other explanation than guilt is reasonably compatible with the circumstances [footnote: See, for example, Martin v Osborne (1936) 55 CLR 367 at 375; Plomp v The Queen (1963) 110 CLR 234 at 243 per Dixon CJ.]. It is of critical importance to recognise, however, that in considering a circumstantial case, all of the circumstances established by the evidence are to be considered and weighed in deciding whether there is an inference consistent with innocence reasonably open on the evidence [footnote: Shepherd v The Queen (1990) 170 CLR 573 at 579 per Dawson J].”

So, because there were circumstances consistent with innocence, the question was whether, in the context of all the evidence in the case there was a reasonable doubt.

This is a slightly different matter from another circumstantial evidence problem that causes much debate: to what standard must circumstantially-proved facts be established before they can be used to support an inference of guilt? In some cases, where the reasoning progresses along a logical path, each step depending on the one before, the answer is clearly that each link in the chain of reasoning must be proved beyond reasonable doubt. But even here, this is not to say that, where more than one item of evidence is needed to constitute proof of the fact relied on for each step, each must be proved beyond reasonable doubt. That is because, where something is proved by circumstantial evidence, the circumstances must be seen in combination, not in isolation. In isolation, they might be doubtful, but the question is whether, in combination, they must be true and, in the case of reasoning in logical steps, each necessary step in the reasoning is established.

It is only the elements of an offence that need to be proved beyond reasonable doubt; other facts need not be.

It is unlikely that there is a difference between the laws of Australia and New Zealand in this regard, although the matter has not always been clear. The High Court of Australia, in the present case at para 48, quoted with approval an earlier decision (the famous “the dingo’s got my baby!” case R v Chamberlain (No 2) (1984) 153 CLR 521 at 535) in which R v Thomas [1972] NZLR 34 at 37-38, 40 was cited. Thomas was a notorious case in New Zealand, the convicted accused ultimately being pardoned. Thomas has also been cited with approval in Canada: R v Morin (1988) 44 CCC (3d) 193.

However, the standard of proof of facts which form the basis for inferences of elements of the offence, may not be universally agreed. The position in Australia was recently summarised by Kirby J, in a case not cited in Hillier: De Gruchy v R (2002) 211 CLR 85 (HCA), at para 47:

“In Australia, but not in England [footnote: Hodge's Case (1838) 2 Lewin 227 [168 ER 1136] per Alderson B; McGreevy v Director of Public Prosecutions [1973] 1 WLR 276; [1973] 1 All ER 503. Samuels, "Circumstantial Evidence", (1986) 150 Justice of the Peace 89] and some other countries, [footnote: As to New Zealand, see R v Hedge [1956] NZLR 511; R v Hart [1986] 2 NZLR 408 at 413; cf Police v Pereira [1977] 1 NZLR 547. As to the position in Canada see R v Cooper [1978] 1 SCR 860 and in the United States see Holland v United States 348 US 121 (1954)] a rather strict approach is taken to the instruction that must be given about circumstantial evidence. The jury must be warned that the primary facts, from which an inference of guilt is to be drawn, must themselves be proved beyond reasonable doubt. The inference of guilt must be the only inference that is reasonably open on all the primary facts which the jury find to be established to the requisite standard of proof. [footnote: Chamberlain v The Queen [No 2] (1984) 153 CLR 521 at 599 per Brennan J; cf R v Van Beelen (1973) 4 SASR 353 at 379-380. See also Peacock v The King (1911) 13 CLR 619 at 634; Glass, "The Insufficiency of Evidence to Raise a Case to Answer", (1981) 55 Australian Law Journal 842 at 852-853.”]"

This is not to say that these primary facts must individually prove guilt beyond reasonable doubt.

Interestingly, in Chamberlain (No 2), and in Thomas, the courts were ultimately shown to have been wrong to uphold convictions, because of errors in the evidence upon which inferences of guilt were based. In Chamberlain (No 2) the red splashes in the car were not, it was later discovered, blood; in Thomas the police were later found to have planted evidence in order to connect the accused with the murder. Perhaps, in now being more rigorous in its requirements concerning the factual basis for inferences, Australia has learnt a lesson that New Zealand still has not.

The third way

For an interesting study in how the most senior judges can differ over problems in criminal law, one need go no further than yesterday’s decision by the High Court of Australia in R v Taufahema [2007] HCA 11 (21 March 2007).

Here, a policeman had been shot dead by a man who got out of a car, of which the respondent, in this Crown appeal, was the driver. There were four occupants in this car, all on parole, all carrying guns, and with them in the car were 2 pairs of gloves and a hockey mask. The car had been, briefly – for up to 20 seconds – chased by the officer in a patrol car, until it came to a halt in an accidental collision. That was when the murderer, Mr P, got out with his gun and killed the victim.

The prosecution’s case had been advanced in two ways at trial. Initially, it was alleged that the men in the car had agreed among themselves to use guns to prevent their being apprehended. By the end of the evidence in the trial, the prosecutor, after a lengthy discussion with the judge, decided to put the case to the jury on the basis that the common agreement between the men was simply to assist each other to avoid apprehension. The advantage this gave the prosecution was that it was unnecessary for the jury to decide that the use of a gun was part of this common agreement. The respondent, Mr T, was convicted of murder as a secondary party, and that must have been because the jury was satisfied that he had foreseen the possibility that, in avoiding arrest, one of the others would use a gun.

Mr T appealed against his conviction to the CCA NSW, on the basis that before he could be liable in this way, a common intent to commit a crime would have to be proved, and here, escaping apprehension was not a crime. Since there was no “foundational” crime, the CCA quashed his conviction and entered an acquittal. The Crown applied, in this present case, to the High Court of Australia for leave to appeal this decision, and for an order for a new trial.

The High Court held, by a majority of 4 to 3, that there should be a new trial. The dissenters were, jointly, Gleeson CJ and Callinan J, and in a separate judgment, Kirby J. The majority (Gummow, Hayne, Heydon and Crennan JJ) delivered a joint judgment.

The Crown’s application was to advance the case on the basis that the men in the car had planned to commit robbery, and this killing was a foreseeable consequence of the pursuit of that common purpose. This way of inviting a conviction had not been offered before in the proceedings.

The main issues were:

•  Should the public interest in the prosecution and conviction of offenders govern the decision?
  
• 
  
•  Should the prosecution be bound by its tactical decisions at trial?
  
• 
  
•  Would ordering a new trial amount to giving the prosecution an opportunity to make a new case, contrary to principles of avoiding double jeopardy?
  
• 
  
•  Should the prosecution be allowed to raise the new point on appeal, when it hadn’t raised it either at trial or on the appeal in the court below?
  
• 
  
•  Was there sufficient evidence of the newly proposed common purpose to warrant a new trial?
  
• 
  
•  How difficult should it be for an appellate court’s order for an acquittal to be overturned?
  
• 
  
•  How much involvement should judges have in the way the prosecution chose to bring its case, especially if the prosecution could be said to have squandered its chances at trial?
  
• 
  
•  Does the jurisdiction of the High Court of Australia include acting as an appellate court or as a court of error? Is the HCA limited to reviewing the exercise of a discretion by the CCA?
  
• 
  
•  Was there a miscarriage of justice for which a more adequate remedy was available than an order for a new trial, or did the fact that this was the murder of a policeman require a decision by a jury rather than an acquittal arising from a deficiency in the law?
  

The case illustrates how important matters of policy can arise from relatively straightforward facts. So many policy issues arose here that any decision could have been justified, which explains the 4 - 3 division between the High Court justices.

The problem with the case at the trial was that the prosecution became conscious of the weakness of the available inference of agreement to use a gun to prevent apprehension, which was the original basis on which it put its case. Instead of accepting almost inevitable defeat, the prosecution changed its stance. It would have been possible to allege, at the outset, liability in the third way - that identified on appeal, involving an allegation of an agreement to commit robbery. Had that been done, the defence would have been fully informed of its jeopardy before the trial.

Sometimes, however, mistakes just can't be allowed to happen.