Melbourne University Law Review
[The ‘proviso’ enables a court hearing a criminal appeal to dismiss the appeal if it accepts that, although there has been some form of error in the trial, there was no ‘substantial miscarriage of justice’. The existence of the proviso reflects the need to balance an accused person’s right to a fair trial, conducted according to law, with the desire to avoid overturning convictions on the basis of inconsequential errors at trial. In this article, I describe and evaluate the High Court’s approach to the proviso and suggest that its application raises a range of difficult issues. I conclude by suggesting that the criminal appeal provision is in need of reform, but that many of the identified difficulties are inherent in the determination of a criminal appeal.]
In each Australian state and territory (apart from the Australian Capital Territory), there is legislation that authorises a court which is hearing a criminal appeal to dismiss that appeal, even where it finds that there has been an error in the trial. Such legislation authorises a court to do so if it considers that the error did not in fact give rise to a ‘substantial miscarriage of justice’. This ‘proviso’ reflects the need to balance an accused person’s right to a fair trial, conducted according to law, with the desire to avoid overturning a conviction on the basis of an inconsequential error at the trial. The question of applying the proviso arises frequently in criminal appeals, yet courts have made few comprehensive attempts to state or analyse the principles underlying the case law.
In this article, I will discuss a range of issues arising from the application of the proviso. I will begin in Part II by briefly summarising the historical background to the proviso and, in Part III, I will examine the competing policy considerations which are central to the application of the proviso. In Part IV, I will consider the relationship between the grounds of appeal and the proviso, and in Part V, I will examine whether the Crown or the appellant bears the onus of proof in relation to the proviso. Part VI will summarise the tests established by the High Court for the application of the proviso: the ‘fundamental error’ test and the ‘lost chance of acquittal’ test. In Parts VII and VIII, I will consider each of these tests in greater detail. Finally, in Part IX, I will consider the prospects for the reform of the proviso.
The proviso has existed for as long as there has been a general right of appeal in criminal cases. No such general right of appeal existed in England or Australia until the 20th century. In the 19th century, there were certain limited procedures for the correction of a legal error. For instance, the ‘case stated’ procedure, established by legislation in England in 1848 and in New South Wales in 1849, enabled the trial judge to refer a point of law for consideration to a higher court. However, there was no basis for appeal if the jury made a wrong factual determination.
The first statute to provide convicted persons with a general right to appeal against conviction, the Criminal Appeal Act 1907, 7 Edw 7, c 23, also introduced the proviso for the first time. This right of appeal, and the accompanying proviso, was first adopted in Australia, in identical terms, by the Criminal Appeal Act 1912 (NSW). It was subsequently adopted by other Australian states between 1912 and 1924.
Each of the criminal appeal provisions adopted a similar wording. For instance, s 6(1) of the Criminal Appeal Act 1912 (NSW) provides that:
The court on any appeal ... against conviction shall allow the appeal if it is of the opinion that the verdict of the jury should be set aside on the ground that it is unreasonable, or cannot be supported, having regard to the evidence, or that the judgment of the court of trial should be set aside on the ground of the wrong decision of any question of law, or that on any other ground whatsoever there was a miscarriage of justice, and in any other case shall dismiss the appeal; provided that the court may, notwithstanding that it is of opinion that the point or points raised by the appeal might be decided in favour of the appellant, dismiss the appeal if it considers that no substantial miscarriage of justice has actually occurred.
The Australian criminal appeal provisions have remained unchanged since their introduction, despite the fact that the English provision on which they were based has, as we shall see, been altered considerably.
In applying the proviso, courts have to take account of a number of competing considerations. As Kirby J said in Festa v The Queen, the standard form criminal appeal provision
seeks at once to uphold the high standards of legal accuracy expected in trials of offenders for criminal offences whilst at the same time recognising that mistakes of varying degrees of significance are difficult or impossible to eliminate completely in any system of criminal justice.
The frequency of dissenting opinions in cases concerning the proviso (which itself has been remarked upon) can, in part, be explained by the difficulty of achieving a balance between these two principles in any particular case.
One theme that has been central to the High Court’s approach is that the proviso must not be allowed to undermine fundamental principles of criminal justice. As Fullagar J said in Mraz v The Queen:
It is very well established that the proviso ... does not mean that a convicted person, on an appeal under the Act, must show that he ought not to have been convicted of anything. It ought to be read, and it has in fact always been read, in the light of the long tradition of the English criminal law that every accused person is entitled to a trial in which the relevant law is correctly explained to the jury and the rules of procedure and evidence are strictly followed. If there is any failure in any of these respects, and the appellant may thereby have lost a chance which was fairly open to him of being acquitted, there is, in the eye of the law, a miscarriage of justice. Justice has miscarried in such cases, because the appellant has not had what the law says that he shall have, and justice is justice according to law.
Australian courts have frequently relied on this passage as authority for the proposition that the test for applying the proviso is a ‘stringent’ one.
Emphasis on the accused’s right to a proper trial will tend to discourage the application of the proviso. Yet, the very existence of the proviso and, in particular, the term ‘substantial miscarriage of justice’, indicates a concern to ensure that convictions are not overturned merely because of a minor and technical error on the part of the trial judge. As Barwick CJ said in Driscoll v The Queen, if
every irregularity of summing up, admission of evidence or in procedure warranted a new trial, the basic intent of the court of criminal appeal provisions would be frustrated and the administration of the criminal law plunged into outworn technicality.
Once a court has allowed an appeal and quashed a conviction, the two remedies available are a retrial or an acquittal. Cases in which the proviso is considered are also, generally, cases in which the appropriate remedy is a fresh trial rather than acquittal. Such a remedy is clearly less drastic than an acquittal — if an appellant is in fact guilty, it can be expected that the second trial will confirm this verdict. Nonetheless, many people would consider that, if a successful appellant was in fact guilty of the crime for which he or she was convicted, a new trial would be undesirable. A new trial could result in an unnecessary use of limited resources (both in terms of the financial cost of a trial and the use of the court’s time) and expose any victims of the crime to greater emotional trauma. Furthermore, considerable practical problems can often be caused by the lapse of time between the events at issue and the new trial.
It is important to note that although the Australian courts of criminal appeal were, at their creation, given the general power to order a retrial, the English Court of Criminal Appeal did not possess such a power until 1988. It is likely that this situation led some judges to either find no error or to apply the proviso more readily where they believed the appellant was guilty, given that the only other alternative would have been to acquit the appellant completely. As Lord Goddard CJ observed to the House of Lords in 1952: ‘I confess that there are cases when one is strongly tempted to apply the proviso, because one very often feels the moral conviction that the man appealing is guilty.’
The terms of the proviso must be considered in the context of the grounds of appeal provided for by the criminal appeal statutes. Each of the appeal provisions states that a court shall allow an appeal against conviction (subject to the proviso) where it is of the opinion
Brooking JA observed in Gallagher that, grammatically, the proviso appears to attach to all three grounds of appeal, and this view seems to be implicit in statements of Isaacs J in Hargan v The King and Barwick CJ in Driscoll. Yet the proviso cannot be applied where the first ground of appeal is made out. As McHugh J observed in TKWJ v The Queen, once the court is satisfied that the verdict should be set aside as unreasonable, or as not supported, having regard to the evidence, it could hardly find that there had been no ‘substantial miscarriage of justice’.
McHugh J also noted that ‘current authority indicates that the proviso has no application in many cases falling within the miscarriage of justice ground.’ In particular, his Honour noted that the proviso will be of no relevance where the miscarriage lies in the unsafe or unsatisfactory nature of the prosecution because, in such a case, the appellant is required to show that the jury should have had reasonable doubt about his or her guilt.
In fact the proviso is not even expressly considered in many cases involving the first and third grounds of appeal. The proviso will usually be relevant in those cases where some specific error is pointed to — for instance, where evidence has been wrongly admitted or where the trial judge misdirected the jury — rather than those cases where the court considers the conviction to be unreasonable.
Where the error in the trial is categorised as a ‘miscarriage of justice’, and the proviso is considered, the court will need to consider the rather elusive distinction between a ‘miscarriage of justice’ and a ‘substantial miscarriage of justice’. The difficulty that this distinction has created in relation to the onus of proof for some appeals based on the miscarriage of justice ground will be considered below.
All these difficulties of interpretation were also observed in relation to the English appeal provision. Partially in response to these criticisms, that provision was substantially amended so as to provide only one ground of review and eliminate the proviso altogether. This reform will be examined in greater detail in the final part of this article, but at this point we can observe that it certainly avoids the problems (identified above) with the relationship between the proviso and the grounds of appeal.
It is well established that, once the appellant has established one of the grounds of appeal, it is for the Crown to satisfy the court that occasion exists for resort to the proviso. This principle, though, must be considered in light of the fact that, as Brooking JA noted in Gallagher, ‘innumerable examples exist of inability to determine whether a judge is applying the proviso or holding that no prima facie case for intervention has been made out.’ If these two issues are not separated, then the onus of proof is unlikely to be split in the way suggested by the authorities.
A particular difficulty arises where there has been an error of fact and the appellant relies on the ‘miscarriage of justice’ ground of appeal. In Simic v The Queen, the High Court observed that an appellant who claims that the trial judge misstated the facts must generally show that there was a miscarriage of justice, whereas an appellant relying on an error of law need simply point to that error. The Court emphasised that there is a ‘fundamental’ distinction between a misdirection of law and of fact. The jury is more likely to be misled by a misdirection of law since they rely on the trial judge for an accurate statement of the law. The jury must form their own view of the facts on the evidence they have heard, so the significance of a misdirection as to fact will vary greatly according to the circumstances.
If, where the error is one of fact, the appellant needs to demonstrate that a miscarriage of justice occurred, does this have the effect of shifting the onus of proof from the Crown to the appellant? As McHugh J observed in TKWJ, one way of reading the miscarriage of justice ground together with the proviso would be to require the appellant to show a ‘miscarriage of justice’ and the Crown to then demonstrate that there had been no ‘substantial miscarriage of justice’. The law has not, though, developed in this way. In Simic, the Court held that the appellant needed to show the misdirection had possibly affected the verdict and that the jury might otherwise have acquitted. This does not appear to leave much room for the Crown to demonstrate that the proviso should be applied.
To further complicate matters, there was a suggestion in Simic that ‘the fact that the case has not been properly presented to the jury will in some circumstances be enough to show that a miscarriage has occurred.’ That is, there will be no need for the appellant to show that the error affected the verdict. This is because ‘an accused person has a fundamental right to a fair trial, conducted according to law.’ This standard is, as will shortly be seen, very similar to that put forward as the basis of the fundamental error test in Wilde v The Queen. This statement in Simic might best be read as suggesting, consistently with Wilde, that where there has been a fundamental error in the trial, the court will allow the appeal without considering whether the accused lost a chance of acquittal.
Since the decision of the High Court in Wilde, it has been clear that there are two separate questions a court must ask in determining whether to apply the proviso.
The first question is whether there has been a fundamental error in the course of the trial (‘the fundamental error test’). The majority in Wilde provided the first comprehensive explanation of this test:
the proviso was not intended to provide, in effect, a retrial before the Court of Criminal Appeal when the proceedings before the primary court have so far miscarried as hardly to be a trial at all. It is one thing to apply the proviso to prevent the administration of the criminal law from being ‘plunged into outworn technicality’ ... it is another to uphold a conviction after a proceeding which is fundamentally flawed, merely because the appeal court is of the opinion that on a proper trial the appellant would inevitably have been convicted. The proviso has no application where an irregularity has occurred which is such a departure from the essential requirements of the law that it goes to the root of the proceedings. If that has occurred, then it can be said, without considering the effect of the irregularity upon the jury’s verdict, that the accused has not had a proper trial and that there has been a substantial miscarriage of justice.
If the error in the appellant’s trial is not considered to be ‘fundamental’, there will still be a question as to whether, due to the error, the appellant has lost a chance of acquittal (‘the lost chance of acquittal test’). The majority in Wilde explained this test as follows:
where there has been a departure from the requirements of a properly conducted trial, it cannot be said that there has been no substantial miscarriage of justice if the applicant has thereby lost ‘a chance which was fairly open to him of being acquitted’ to use the phrase of Fullagar J in Mraz v The Queen or ‘a real chance of acquittal’ to use the phrase of Barwick CJ in R v Storey.
In this part, I will consider the fundamental error test in some detail. I will begin by considering why the lost chance of acquittal test would be inappropriate where a fundamental error has occurred. I will suggest that the concern with fundamental errors reflects a focus on ‘due process’, as distinct from a concern with the appellant’s guilt or innocence. In the second section, I will consider what type of errors have been viewed by the courts as ‘fundamental’. I will suggest that this doctrine is rarely applied and that there is little in the way of guidance to determine whether a particular error is fundamental. In the third section, I will argue that the decision in Wilde reveals a tension between the principle of due process and that of truth. The reasoning of the majority in Wilde suggests that, where the question of the appellant’s guilt or innocence, rather than due process, becomes the focus of the decision, there is little difference between the fundamental error test and the lost chance of acquittal test.
The High Court in Wilde stated that the lost chance of acquittal test was inadequate in situations where there had been a fundamental error in the course of the trial. Why was this so? Many people might wonder why the role of a criminal appeal court is not simply to determine what we might label the ‘truth’ — that is, whether the appellant is in fact guilty or innocent. If it is possible to say this, then why must the appeal be allowed, with the consequence of an acquittal or a new trial? To understand the High Court’s approach to ‘fundamental’ errors, it is necessary to appreciate that the primary question — as suggested by the wording of the proviso itself — is not whether the appellant is or is not guilty, but rather whether a ‘substantial miscarriage of justice’ occurred in the appellant’s trial. As Lloyd LJ said in respect of the equivalent English provision, ‘nothing in [the appeal provision], or anywhere else obliges or entitles us to say whether we think the appellant is innocent.’
How can there be a ‘substantial miscarriage of justice’ if a person who is in fact guilty is convicted? The answer is that there may have been a failure of due process — that is, the process of conviction may not have been fair — and a criminal appeal court also performs the important task of ensuring that this requirement is observed. As McBarnet has observed, criminal courts perform dual functions: ‘[The judge] must not just ensure that justice is done in the sense of the accused getting his deserts; they must also ensure that the technical checks on how justice is executed are upheld.’
An appeal court needs to preserve the rights of suspects during investigation, and the right of defendants to a fair trial, even if the particular suspect or defendant in the case at hand is in fact guilty. The court cannot focus solely on the particular appellant because the legal rulings that it makes will apply in future cases. The principle is best illustrated by examples of an extreme failure of due process. Suppose that a confession was extracted from a defendant by torturing him or her; or that a defendant was not allowed to present a defence at his or her trial. Is this not such a serious error that, even if the appeal court knew (by some independent means) that the appellant was guilty, it would be appropriate to order a new trial?
In the United Kingdom and the United States, it has been widely recognised that the tension between these principles of ‘truth’ and ‘due process’ is important in the context of criminal appeals. There has been no such explicit recognition of the conflict between these principles in this context in Australia, although, as I will suggest below, an understanding of them does provide one way to analyse the decision in Wilde.
The values of ‘due process’ and ‘truth’ will often coexist and indeed promote each other. Most obviously, a fair trial is the best way to know whether a person is in fact guilty. However, these values will on occasion conflict and an adherence to one over the other will produce a different outcome in a criminal appeal. If ‘truth’ is the ultimate value, then an appellant who is in fact guilty must not be acquitted, regardless of the severity of the error that occurred in his or her trial. By contrast, ‘due process’ would require that, once a sufficiently serious error was made, the appeal be allowed, regardless of whether the appellant is guilty or innocent.
On one view, an appeal court could resolve this dilemma by not applying the proviso and ordering a retrial. Assuming that a guilty appellant will be convicted following a retrial, this option ‘simultaneously protects the right to due process and ensures that the appellant gets his just deserts.’ There were some predictions in the United Kingdom that, once the Court of Criminal Appeal had the power to order a new trial, the proviso would wither away. But there is no evidence that this has happened. As the discussion in Part III showed, a retrial is not a cost-free option. As such, courts have sought to confine the cases in which they allow an appeal (and, generally, order a new trial) to those where there is a real doubt about the appellant’s guilt or there has been a more significant failure of due process. Hence the ‘fundamental error’ test. But, as the next section will show, there is no easy way to distinguish between ‘fundamental’ and other errors.
The above analysis showed why the High Court would wish to treat fundamental errors differently. But what is a fundamental error? The majority in Wilde made it clear that:
There is no rigid formula to determine what constitutes such a radical or fundamental error. It may go either to the form of the trial or the manner in which it was conducted ... the wording of the proviso is quite general and it is clear that it may be applied notwithstanding a misdirection concerning the law or the wrongful admission of evidence. In the end no mechanical approach can be adopted and each case must be determined upon its own circumstances.
An examination of the case law suggests that not only is there no ‘rigid formula’, but also that it is not possible to state any guiding principles as to what constitutes a fundamental error. Despite the distinction drawn in Wilde, the Court has not generally devoted a great deal of attention to the differences between those errors which are ‘fundamental’ and those which are not. Examples of disagreement as to whether an error is in fact fundamental, or is merely such as to suggest that the appellant lost a reasonable chance of acquittal, are illustrative of the absence of guiding principles in this area. The fact that many appellants, when faced with the application of the proviso, will present both branches of the Wilde test as alternative arguments is also reflective of this lack of a clear distinction.
A similar observation has been made in relation to the English Court of Criminal Appeal:
There does not seem from the judgments to be any clear explanation laid down of what faults are so fundamental or cardinal that, irrespective of how much other material there is to justify a conviction ... the proviso cannot be upheld ... The line between some of these examples and some of the cases tabulated as really serious faults where the proviso has been exercised ... seems narrow and blurred.
Green v The Queen provides a good illustration of the extent to which a court can disagree as to whether an error was fundamental. The error in Green concerned the failure of the trial judge to direct the jury to have regard to the appellant’s family circumstances in considering whether a defence of provocation was available. Brennan CJ said that Wilde
applies only where there is ‘such a departure from the essential requirements of the law that it goes to the root of the proceedings’ so that ‘the accused has not had a proper trial.’ It applies only to fundamental irregularities which demonstrate that no proper trial has taken place. It does not apply when there is no more than an erroneous ruling on the admissibility of evidence or a misdirection on a particular point of fact or law arising in the trial.
Brennan CJ specifically stated that he did not regard the error in this case as fundamental, though he did think that the proviso should be applied on the basis that the appellant had lost a chance of acquittal. McHugh J, by contrast, considered that the errors in question led to the trial being fundamentally flawed because the ‘accused’s real case on provocation was never left to the jury.’ Gummow and Kirby JJ concluded that there had been no substantial miscarriage of justice at all. Kirby J considered that the ‘technical’ mistakes which occurred in the trial certainly did not constitute fundamental errors.
In addition to disagreement within the one case, there are also instances where similar errors are seen to be fundamental in one case, but not in another. In S v The Queen, a majority of the High Court concluded that a trial in which the accused had not known with certainty the precise charge he was to meet was fundamentally flawed. The accused had been charged with incest. Although the complainant had given a general account of sexual abuse, no particular act or acts were identified as the subject of the offences charged in the indictment.
In contrast, in KBT v The Queen, where a similar issue arose, a majority of the Court applied the alternative test, namely whether the accused was deprived of a chance of acquittal that was fairly open. The accused in that case was charged with the offence of maintaining an unlawful sexual relationship with a child, which was defined as committing an offence of a sexual nature on three or more occasions. The complainant gave evidence of various incidents and the trial judge erred in not directing the jury that they needed to agree that the same three incidents amounted to offences of a sexual nature.
Much of the guidance provided by the High Court has come in the form of establishing what does not necessarily constitute a fundamental error. Wilde does not itself provide an illustration of a fundamental error, since the majority concluded that there was no fundamental error on the facts of that case. The majority observed that a defect in the form of the indictment — in the sense that the appellant should have been tried separately for offences joined together in one indictment — does not necessarily lead to a mistrial.
In Glennon v The Queen, a majority of the High Court concluded that ‘[a]lthough the right to silence is a fundamental right of any accused person, it cannot be said that any misdirection on that subject is a fundamental irregularity of the kind discussed in Wilde.’ Deane and Gaudron JJ observed that ‘[a] misdirection as to the use or evaluation by a jury of properly admitted evidence does not ordinarily result in there not being, overall, a fair trial according to the law in the sense discussed ... in Wilde.’
In Krakouer v The Queen, a majority of the High Court rejected the appellant’s contention that there was a fundamental error because the trial judge gave a misdirection as to proof of an element of the offence with which the appellant was charged. The Court in Krakouer did, though, go on to find that the appellant had lost a reasonable chance of acquittal.
The most definitive suggestion by the High Court as to something that does necessarily constitute a fundamental error predates the decision in Wilde. In Quartermaine v The Queen, Gibbs J drew on the earlier decision of Andrews v The Queen as demonstrating that ‘the proviso should not be applied where the accused was not in reality tried for the offences for which he was indicted.’ In both Quartermaine and Andrews, the High Court declined to apply the proviso on the basis that the summing up delivered by the trial judge was more appropriate to a different charge and the jury was not instructed as to the essential elements of the charge in fact laid.
An understanding of the tension between the values we have labelled ‘truth’ and ‘due process’ provides a useful basis to analyse the decision in Wilde. I would suggest that the majority in Wilde focused on ‘truth’ rather than on ‘due process’. Because of this, there was in fact little difference between their application of the fundamental error test and the lost chance of acquittal test. By contrast, Deane and Gaudron JJ do appear to have focused on due process, with the consequence that they did not examine whether the appellant in Wilde was in fact guilty.
The issue in Wilde was whether there had been a miscarriage of justice as a result of the applicant not receiving, as he should have, separate trials for two different offences. A further error lay in the trial judge’s directions on similar fact evidence in relation to the two offences.
Brennan, Toohey and Dawson JJ held that it was appropriate to apply the proviso and dismissed the appeal. Their Honours took the view that the relative strength of the prosecution and defence cases was relevant to the question of whether a ‘fundamental error’ had occurred. The purpose of weighing up the relative strength of the prosecution and defence cases where a ‘fundamental error’ is alleged is to determine the ‘gravity and significance’ of the error, rather than whether the applicant would inevitably have been convicted notwithstanding the error. Here, the relative strength of the prosecution case meant that there had been no fundamental error.
It is difficult to see how, for practical purposes, it makes any difference whether the relative strength of these opposing cases is considered so as to determine whether there was a fundamental error or whether the appellant lost a reasonable chance of acquittal. The end result is that an appellant cannot succeed where there has been, in spite of the identified error, a sufficiently strong prosecution case — as there was in Wilde. Despite the majority’s rhetorical emphasis on ‘fundamental’ errors in their reasoning in Wilde, this suggests that their primary concern was with ‘truth’ rather than ‘due process’.
The approach of the majority in Wilde can be contrasted with that of Deane J who (together with Gaudron J) dissented. Deane J said that
the applicant’s trial was so pervaded by unfairness and error ... that the conclusion is inescapable that the applicant was denied a fair trial according to law of those alleged offences. That being so, it is not to the point that the case against him ... appears, at this distance from the impact of live evidence and the atmosphere of the trial, to have been an overwhelmingly strong one.
Deane J’s approach is closer to the ‘due process’ model. Although his Honour appeared to share the majority’s view as to the strength of the prosecution case, he considered that the nature of the error required that the applicant receive a new trial. The question of whether the applicant was in fact guilty was, for Deane J, irrelevant.
In determining whether the appellant has lost a real chance of acquittal, the court will consider the strength of the prosecution case and the significance of the error. It must make its own ‘assessment of the facts of the case’. Barwick CJ said in Driscoll that:
It is for the court itself to be affirmatively satisfied [that there has been a substantial miscarriage of justice], and for this purpose the court will consider for itself the evidence and the inferences properly available therefrom. Where the credit to be given to oral evidence is in question, the court will act on that view which the court thinks the jury must have formed, having regard to the verdict they have returned or, where no inference can be relevantly drawn from the verdict, the view which the court thinks they could reasonably have formed.
The need for the appeal court to revisit the evidence presents significant practical difficulties. As Deane J said in Dietrich v The Queen, this review effectively involves ‘a trial of an accused person by appellate judges who [have] seen no witnesses, heard no evidence and [have] had no direct contact with the atmosphere, the tensions, the nuances or the reality of the actual trial.’
The need to avoid a ‘trial’ by the appeal court, rather than the jury, has long been a major concern in relation to criminal appeals. For many years, the English Court of Criminal Appeal justified a restrictive interpretation of the right to appeal because of its concern not to interfere with the constitutional role of the jury. This same concern was one of the key reasons that the ‘fundamental error’ test was developed.
Yet, even where there has not been a fundamental error, it will often prove difficult or impossible for the appeal court to put themselves in the position of the jury so as to assess the evidence against the appellant in order to determine whether he or she lost a reasonable chance of acquittal. Since the reasoning of the jury is not publicly available, it is impossible to calculate with certainty the effect that the error would have had on their ultimate finding of guilt. This is why the appeal court will often have no choice but to form its own view of the facts, however unsatisfactory such a procedure may be.
In Festa, McHugh J said that:
The prevalence of dissenting views in cases dealing with the application of the proviso illustrates the largely subjective nature of the inquiry, resting as it does on factors such as the error alleged, the relative strength of the prosecution and defence cases and the court’s characterisation of the hypothetical jury, ‘acting reasonably’ and properly directed.
A consideration of some of the factors relevant to determining whether an accused has lost a chance of acquittal demonstrates that, as with the fundamental error test, there are few guiding principles in this area. The decisions suggest that, in every case, the question of whether the appellant did lose a reasonable chance of acquittal will depend on an individual interpretation of the particular facts of the case.
It is clear that appeal courts can, and frequently do, conclude that evidence which was not tainted by any error is sufficient to sustain the appellant’s conviction. However, care must be taken in this regard.
In Domican v The Queen, Brennan J cautioned that if there were two independent bases on which it was possible for the jury to convict, and the misdirection or inadequate direction of the trial judge related to only one of these bases, the appeal court cannot assume that the jury relied on the basis which was unaffected. In Domican, there were three distinct categories of evidence, any one of which might have entitled the jury to find the appellant guilty of shooting with intent to kill. The trial judge erred in directing the jury as to the identification evidence which constituted one of these three categories. Brennan J explained that, since it was possible that the jury verdict might have been based on the identification evidence, it could not be assumed that the appellant had not lost a reasonable chance of acquittal.
The majority in Domican took the view that the proviso could only have been applied if the jury must inevitably have convicted the accused, independently of the identification evidence. There were potential weaknesses in the remaining evidence against Domican which led the Court to decline to apply the proviso.
Domican was distinguished in Festa on the basis that the untainted evidence in Festa made the appellant’s conviction inevitable.
In KBT, a majority of the High Court said:
There are occasions when [the proviso] is properly applied where a point was not taken at the trial because, for example, it was not in issue or there was some forensic advantage to be gained by not raising it. In cases of that kind, the provision is applied because, having regard to the defence case, the accused was not deprived of a chance of acquittal that was fairly open ... if the appellant was deprived of a chance of that kind, the fact that no complaint was made at trial is irrelevant.
In Doggett v The Queen, Kirby J concluded that the failure by experienced counsel to raise an objection to the trial judge’s failure to give a specific warning about the danger of convicting on the evidence of the complainant in light of delay could not have been a matter of strategy, and so did not preclude the application of the proviso.
In De Jesus v The Queen, Mason and Deane JJ (in the minority) concluded that there had not been a substantial miscarriage of justice in circumstances where the appellant’s counsel had failed to inform the trial judge of the factual matter to which, it was then argued on appeal, the judge had accorded insufficient weight.
Sometimes, the appeal court will use the fact that the appellant was acquitted on one count as some indication of doubt on the part of the jury in relation to the other counts. In Dietrich, the trial judge erred in not allowing an adjournment so that the accused could obtain counsel. Mason CJ and McHugh J said that the ‘not guilty’ verdict returned by the jury on one of the counts was central to their conclusion that the appellant had lost a chance of acquittal. Their Honours observed that:
The evidence against the applicant appears strong on all counts but, in circumstances where the jury found him not guilty on one count, how can this Court conclude that, even with the benefit of counsel, the applicant did not have any prospect of acquittal on [the other count], of which he was then deprived by being forced to trial unrepresented?
In Glennon, a majority of the Court considered that the acquittal of the appellant by the jury in relation to two of the counts reinforced the conclusion that the error of the trial judge had deprived the appellant of a chance of acquittal in relation to other counts.
However, the relevance of acquittal on some counts to the application of the proviso will depend greatly on the precise facts of the case. There may be reasons why the jury acquitted on some counts, which in no way reflect on their verdicts in relation to other counts. In Farrell, the appellant had been charged with a number of counts relating to sexual assault of the complainant. The counts on which the appellant was acquitted were those in which he was charged with aiding and abetting another person to assault the complainant, rather than as the principal offender. Hayne J noted that, in the particular circumstances of the case, he attached ‘no weight to the different verdicts given by the jury to the various counts charged.’ On the evidence, it was understandable that the jury was not convinced beyond reasonable doubt that the appellant had actively encouraged the other offender to assault the complainant. The acquittal on these counts did not, though, cast any doubt on the jury’s convictions of the appellant as the principal offender.
In Farrell, the Justices of the High Court disagreed as to whether other directions given by the trial judge had the effect of overcoming his Honour’s error in directing the jury that the opinion of a psychiatrist as to the complainant’s mental condition was irrelevant. McHugh and Hayne JJ held that this error did not have the effect of disadvantaging the appellant since the trial judge nonetheless warned the jury of the need to take special care in accepting the complainant’s evidence in light of his mental problems. Gaudron, Kirby and Callinan JJ each considered that the error had amounted to a substantial miscarriage, in spite of the other directions given by the trial judge.
This article has demonstrated that courts are faced with many difficult and unclear issues when applying the proviso in the context of a criminal appeal. The relationship between the proviso and the grounds of appeal is in many respects unsatisfactory and in particular creates confusion as to the onus of proof where the appellant relies on the miscarriage of justice ground in relation to a misdirection on the facts.
In light of such issues, the question arises whether the criminal appeal provision would benefit from reform. As has been noted, the English criminal appeal provision on which the Australian provisions were originally modelled has been amended on a number of occasions. Section 2 of the Criminal Appeal Act 1995 (UK) c 35 states that the Court of Appeal
One significant aspect of this new provision — the replacement of three grounds of appeal by one — is not the focus of this article. The issue for present purposes is whether the removal of the proviso, with its separate reference to ‘substantial miscarriage of justice’, represents an improvement to the criminal appeal provision. The current English provision clearly has the advantage of simplicity when compared with its poorly expressed Australian counterparts. In particular, as noted above, this formulation would avoid the problems associated with the interaction between the grounds of appeal and the proviso. This is, I would suggest, reason alone for a similar reform in Australia.
It is to be hoped that, at some point in the near future, there will be a review of the criminal appeal provisions, similar to that which has occurred on a number of occasions in the United Kingdom. Such an inquiry could consider a simplification of the criminal appeal provisions. It could also provide a forum in which the interpretative principles relevant to the proviso could be addressed more comprehensively than in the current situation, where courts consider these issues in a somewhat piecemeal fashion.
However, we could not expect that reformulating the criminal appeal provision would resolve many of the issues raised in this article. The experience in the United Kingdom has been that changes to the wording of the appeal provision do little to alter the approach of the courts. The 1995 amendment, although significant on its face, was primarily intended to maintain the essence of the earlier law while simplifying the form. I would suggest that the two difficulties associated with the application of the proviso, and identified in this article (and summarised below), are inherent in the resolution of a criminal appeal.
First, the court will often have to resolve difficult factual issues to determine whether the appellant’s conviction can be upheld in spite of an error at trial. Appeal courts have always felt uneasy about being required to determine whether an appellant is in fact guilty, a task they see as much better suited to the trial court. As long as there exists a general right of appeal against criminal conviction, but it is deemed undesirable that a minor and inconsequential error will lead to the quashing of a conviction, the courts will continue to be required to perform this difficult task. Nor is there any avoiding the reality that the question of whether an appellant lost a reasonable chance of acquittal is one on which reasonable people could differ.
Secondly, there are the difficulties posed by the dual function of a criminal appeal court — to find out the ‘truth’ of the particular case and to maintain principles of due process. The proviso plays a central part in allowing courts to balance these roles. A cynic might suggest that the proviso enables courts to maintain the rhetoric of due process while ensuring that the convictions of guilty persons are not reversed. The court can condemn the error that occurred in the trial, hence performing its general role of upholding due process, yet confirm the conviction, ensuring the guilty individual in question receives his or her just deserts.
But is rhetorical condemnation of a failure of due process enough? Certainly the ‘fundamental error’ test suggests that the High Court believes that the proviso should not be used in this way, at least where the failure of due process was sufficiently serious. But is the fundamental error doctrine itself more a matter of rhetoric than of substance — given that it is rarely applied, of uncertain nature and, most importantly, in its application still focuses on the strength of the case against the appellant?
The tension between these dual roles of a criminal appeal court, and the question of the relative weight to be given to each, are issues that would benefit from more explicit consideration. Given that the issue is largely one of reconciling conflicting policy objectives, this is a role that might best be performed by either a committee of inquiry, along the lines of the English committees referred to above, or a law reform body.
[∗] BA, LLB (Hons) (ANU). I would like to thank Edward Santow and the anonymous referees for their comments.
 Criminal Appeal Act 1912 (NSW) s 6(1); Criminal Code (NT) s 411(2); Criminal Code 1899 (Qld) s 668E(1A); Criminal Law Consolidation Act 1935 (SA) s 353(1); Criminal Code (Tas) s 402(2); Crimes Act 1958 (Vic) s 568(1); Criminal Code (WA) s 689(1). These various pieces of legislation only apply to appeals to the courts of criminal appeal (which are, in some jurisdictions, called the Court of Appeal) and not to any other criminal appeals, such as an appeal from the decision of a Magistrate to the District or County Court or Supreme Court. As to the position in the Australian Capital Territory, see Conway v The Queen (2002) 209 CLR 203 and Supreme Court Act 1933 (ACT) s 37O.
 There does not appear to have been any comprehensive empirical analysis in the Australian context of the application of the proviso, although Brooking JA suggested in R v Gallagher  2 VR 671, 673 (‘Gallagher’) that the issue arises ‘almost ... daily’. In relation to the United Kingdom, see Kate Malleson, ‘Review of the Appeal Process: Research Study No 17’ in United Kingdom, Royal Commission on Criminal Justice, The Runciman Report (1993) 12–13 (‘Runciman Report’); Michael Knight, Criminal Appeals (1970) 15–21.
 See Conway v The Queen (2002) 209 CLR 203, 208–13 (Gaudron ACJ, McHugh, Hayne and Callinan JJ); Sir James Stephen, A History of the Criminal Law of England (1883) 308–12; Richard Nobles and David Schiff, Understanding Miscarriages of Justice (2000) 16–17.
 Other ways of challenging a conviction included the writ of error and seeking an executive pardon: see G D Woods, A History of Criminal Law in New South Wales: The Colonial Period 1788–1900 (2002) 253–5; Rosemary Pattenden, English Criminal Appeals 1844–1994 (1996)
 The Canadian criminal appeal provision, Criminal Code, RSC 1985, C-46, s 686, is also modelled on the English provision.
 (2001) 208 CLR 593, 653 (‘Festa’).
 Ibid; Grey v The Queen  HCA 65; (2001) 184 ALR 593, 603 (Gleeson CJ, Gummow and Callinan JJ).
  HCA 59; (1955) 93 CLR 493, 514.
 Zoneff v The Queen (2000) 200 CLR 234, 267 (Gleeson CJ, Gaudron, Gummow and Callinan JJ); Farrell v The Queen (1998) 194 CLR 286, 326 (Callinan J) (‘Farrell’).
  HCA 43; (1977) 137 CLR 517, 527 (‘Driscoll’). See also R v Thompson (1913) 9 Cr App R 252, 260 (Isaacs CJ).
 See, eg, Criminal Appeal Act 1912 (NSW) ss 6(2), 8; Crimes Act 1958 (Vic) s 568(2).
 In the great majority of cases where the proviso is considered, but not applied, a new trial is ordered. This is probably because, in such cases, there is sufficient evidence to justify a conviction; see DPP (Nauru) v Fowler  HCA 48; (1984) 154 CLR 627, 630 (Gibbs CJ, Murphy, Wilson, Deane and Dawson JJ); Andrews v The Queen  HCA 84; (1968) 126 CLR 198, 211 (Barwick CJ, McTiernan, Taylor, Windeyer and Owen JJ).
 The problems associated with the outright acquittal of persons who are in fact guilty are discussed in A L Goodhart, ‘Acquitting the Guilty’ (1954) 70 Law Quarterly Review 514.
 A belief in ‘due process’, as discussed below, might lead some people to reject this view. For instance, in Runciman Report, above n 2, 216, only a ‘majority’ of the Commission members took the view that ‘if the error would have made no difference to the safety of the conviction, the appeal should be dismissed’.
 See, eg, Criminal Appeal Act 1912 (NSW) s 8.
 The Criminal Justice Act 1988 (UK) c 33 removed the requirement that the appeal be based on fresh evidence, as had been the case under the Criminal Appeal Act 1968 (UK) c 19: see Pattenden, above n 4, 198.
 J R Spencer, ‘Criminal Law and Criminal Appeals — The Tail That Wags the Dog’  Criminal Law Review 260, 261.
 United Kingdom, Parliamentary Debates, House of Lords, 8 May 1952, 793, quoted in Pattenden, above n 4, 189.
  2 VR 671, 675.
  HCA 45; (1919) 27 CLR 13, 23.
  HCA 43; (1977) 137 CLR 517, 524.
 (2002) 193 ALJR 7, 19 (‘TKWJ’).
 Ibid 18.
 Ibid 17–18. See also M v The Queen  HCA 63; (1994) 181 CLR 487, 494–5 (Mason CJ, Deane, Dawson and Toohey JJ); R v Maguire  QB 936, 951–2.
 See, eg, Chamberlain v The Queen [No 2]  HCA 13; (1984) 153 CLR 514; Morris v The Queen  HCA 50; (1987) 163 CLR 454; M v The Queen  HCA 63; (1994) 181 CLR 487.
 See Gallagher  2 VR 671, 675–9 (Brooking JA).
 J C Smith, ‘The Criminal Appeal Act 1995: (1) Appeals against Conviction’  Criminal Law Review 920. See also Pattenden, above n 4, 189–90.
 Criminal Appeal Act 1995 (UK) c 35, s 2.
 Cohen and Bateman (1909) 2 Cr App R 197, 207–8; Mraz v The Queen  HCA 59; (1955) 93 CLR 493, 514 (Fullagar J); Driscoll  HCA 43; (1977) 137 CLR 517, 524–5 (Barwick CJ).
  2 VR 671, 674. See also TKWJ  HCA 46; (2002) 193 ALR 7, 19–20 (McHugh J).
  HCA 25; (1980) 144 CLR 319, 331 (‘Simic’).
 Ibid 331–2.
  HCA 46; (2002) 193 ALR 7, 23–4. See also Gallagher  2 VR 671, 679 (Brooking JA).
  HCA 25; (1980) 144 CLR 319, 332.
 Ibid 331 (emphasis added).
  HCA 6; (1988) 164 CLR 365 (‘Wilde’).
 Ibid 373 (Brennan CJ, Dawson and Toohey JJ). The majority adopted the statement of Gibbs J in Quartermaine v The Queen  HCA 29; (1980) 143 CLR 595, 600–1. Although the majority in Wilde do not refer to them, there are also decisions of the English Court of Criminal Appeal which clearly support this reasoning: see R v McKenna (1960) 44 Cr App R 63; R v Oliva (1960) 46 Cr App R 241. Similarly, Michael Knight draws a distinction between a sufficiently serious error and an error that is not sufficiently serious: see Michael Knight, ‘Convicting the Guilty’  Criminal Law Review 24, 26–7.
  HCA 6; (1988) 164 CLR 365, 371 (Brennan CJ, Dawson and Toohey JJ) (citations omitted).
 Ibid 373.
 The question of what it means to say a person is ‘guilty’ is itself a complex one. The High Court recently confirmed, in Eastman v DPP (ACT) (2003) 198 ALR 1, that the ordinary meaning of ‘guilt’ is guilt as properly established in a court of law. A person is not ‘guilty’ just because the elements that constitute the offence against him or her exist. Thus, a doubt as to whether Eastman was fit to plead was a doubt as to his ‘guilt’.
 R v McIlkenny (1991) 93 Cr App R 287, 311. The principle is also illustrated by a comment made by Lord Goddard CJ to a successful appellant: ‘Do not think that we are doing this because we think that you are an innocent man. We do not. We think that you are a scoundrel’: ‘Inadmissibility of Evidence of Former Wife: “Scoundrel” Wins Appeal’, The Times (London), 17 November 1953, 2.
 This proposition draws on the analysis in Nobles and Schiff, above n 3, 16–17. In criminal justice more generally, a distinction has been identified between ‘crime control’ and ‘due process’: see Herbert Packer, The Limits of the Criminal Sanction (1968) 149–73.
 Doreen McBarnet, Conviction: Law, the State and the Construction of Justice (1981) 158 (emphasis in original).
 See Australian Law Reform Commission, Principled Regulation: Federal Civil and Administrative Penalties in Australia, Report No 95 (2002) [14.5]–[14.9]; Packer, above n 43, 163–73.
 Pattenden, above n 4, 57–8, 70–1, 165, 207–10; McBarnet, above n 44, 157–9. See also David Schiff and Richard Nobles, ‘Criminal Appeal Act 1995: The Semantics of Jurisdiction’ (1996) 59 Modern Law Review 573; David Schiff and Richard Nobles, ‘Due Process and Dirty Harry Dilemmas: Criminal Appeals and the Human Rights Act’ (2001) 64 Modern Law Review 911.
 Packer, above n 43, 227–38.
 Pattenden, above n 4, 189.
 Ibid 188–9.
  HCA 6; (1988) 164 CLR 365, 373 (Brennan CJ, Dawson and Toohey JJ).
 See, eg, Murray v The Queen  HCA 26; (2002) 211 CLR 193. The trial judge erred in directing the jury as to the onus of proof — a matter which might well be seen as a fundamental error — yet none of the judges classified the error as fundamental or otherwise.
 See, eg, Glennon v The Queen (1994) 179 CLR 1, 7 (Mason CJ, Brennan and Toohey JJ); Krakouer v The Queen  HCA 43; (1998) 194 CLR 202, 211–12 (Gaudron ACJ, Gummow, Kirby and Hayne JJ).
 Knight, Criminal Appeals, above n 2, 10.
 (1997) 191 CLR 334 (‘Green’).
 Ibid 346–7 (citations omitted).
 Ibid 372.
 Ibid 387 (Gummow J), 413 (Kirby J).
 Ibid 413.
  HCA 66; (1989) 168 CLR 266, 277–8 (Dawson J), 281–2 (Toohey J), 287–8 (Gaudron and McHugh JJ).
 (1997) 191 CLR 417, 423–4 (Brennan CJ, Toohey, Gaudron and Gummow JJ) (‘KBT’).
 Wilde  HCA 6; (1988) 164 CLR 365, 373 (Brennan CJ, Dawson and Toohey JJ).
 (1994) 179 CLR 1, 8 (Mason CJ, Brennan and Toohey JJ) (‘Glennon’).
 Ibid 12.
  HCA 43; (1998) 194 CLR 202, 212 (Gaudron, Gummow, Kirby and Hayne JJ) (‘Krakouer’).
  HCA 29; (1980) 143 CLR 595 (‘Quartermaine’).
  HCA 84; (1968) 126 CLR 198, 207–10 (Barwick CJ, McTiernan, Taylor, Windeyer and Owen JJ) (‘Andrews’).
  HCA 29; (1980) 143 CLR 595, 601.
 Wilde  HCA 6; (1988) 164 CLR 365, 374–5.
 Ibid 374 (Brennan CJ, Dawson and Toohey JJ).
 Wilde  HCA 6; (1988) 164 CLR 365, 376.
 Driscoll  HCA 43; (1977) 137 CLR 517, 525 (Brennan CJ); Simic  HCA 25; (1980) 144 CLR 319, 332.
 Wilde  HCA 6; (1988) 164 CLR 365, 372 (Brennan CJ, Dawson and Toohey JJ).
  HCA 43; (1977) 137 CLR 517, 525.
  HCA 57; (1992) 177 CLR 292, 338 (‘Dietrich’).
 R v McGrath  2 All ER 495, 496 (Lord Goddard CJ); Pattenden, above n 4, 141–2.
 Quartermaine  HCA 29; (1980) 143 CLR 595, 601 (Gibbs J); Wilde  HCA 6; (1988) 164 CLR 365, 382, 385 (Gaudron J); Makin v A-G (NSW)  UKPC 56;  AC 57, 69–70 (Lord Herschell LC).
 (2001) 208 CLR 593, 631–2. His Honour cites as examples of cases where dissent has occurred R v Miller (1980) 25 SASR 170; Liberato v The Queen  HCA 66; (1985) 159 CLR 507; Green (1997) 191 CLR 334; Farrell (1998) 194 CLR 286.
 See, eg, Festa (2001) 208 CLR 593; Conway v The Queen (2002) 209 CLR 203; Harriman v The Queen  HCA 50; (1989) 167 CLR 590.
 (1992) 173 CLR 555, 570–1 (‘Domican’) (Brennan J). Although Brennan J dissented as to whether there had been an error at trial, he agreed with the majority that if there was an error, the proviso could not be applied in this case: ibid.
 Ibid 571.
 Ibid 565–6 (Mason CJ, Deane, Dawson, Toohey, Gaudron and McHugh JJ).
 (2001) 208 CLR 593, esp 655–7 (Kirby J).
 (1997) 191 CLR 417, 423–4 (Brennan CJ, Toohey, Gaudron and Gummow JJ). See also Conway v The Queen (2002) 209 CLR 203, 241–2 (Kirby J); Krakouer  HCA 43; (1998) 194 CLR 202, 216 (Gaudron, Gummow, Kirby and Hayne JJ).
 (2001) 208 CLR 343, 382–3.
  HCA 65; (1986) 68 ALR 1, 10–11.
 Dietrich  HCA 57; (1992) 177 CLR 292, 314–15.
 Ibid (citations omitted).
 (1994) 179 CLR 1, 9 (Mason CJ, Brennan and Toohey JJ).
 (1998) 194 CLR 286.
 Ibid 311.
 (1998) 194 CLR 286, 310 (Hayne J, McHugh J agreeing).
 Ibid 293–4 (Gaudron J), 301 (Kirby J), 326–7 (Callinan J).
 The English proviso had earlier been amended by the Criminal Appeal Act 1968 (UK) c 19, s 4(1)(c) to refer to a ‘miscarriage of justice’ rather than a substantial miscarriage of justice.
 See Interdepartmental Committee on the Court of Criminal Appeal, United Kingdom, The Donovan Report (1965); Runciman Report, above n 2.
 Nobles and Schiff, above n 3, 82–9.
 R v Chalkley  EWCA Crim 3416;  QB 848, 859 (Auld LJ); R v Mullen  EWCA Crim 36;  QB 520, 540 (Rose LJ).
 This suggestion draws on McBarnet, above n 44, 159–62.