Melbourne v The Queen

3. Roy Bernard Melbourne (the accused) was convicted of murder in the Supreme Court of the Northern Territory on 13 June 1996 after a trial before a jury and Thomas J. The Court of Criminal Appeal of the Northern Territory (Martin CJ, Gallop and Angel JJ) dismissed his appeal against the conviction. Pursuant to a grant of special leave, he now appeals to this Court.

4. In July 1995, the accused, who had lived next door to the deceased, Mrs Irene Chambers, for about ten months, killed her by stabbing her three times. At the scene, he told [1] one person, in a very flat tone of delivery, that he "did it because she just wouldn't stop". When police officers arrived at the scene, they observed that the accused was "vacant" and "dazed or intoxicated or both". He did not respond to their questions and was unsteady on his feet. The accused was breathalysed and found to have a blood alcohol content of 0.136 per cent. He was seen by a doctor some hours after being taken into custody and said that he did not know why he had been arrested. In an interview the next day, a police officer asked the accused whether he had any recollection of stabbing the deceased. The accused replied "None - none whatsoever".

1. The jury were entitled to conclude from other evidence tendered at the trial that this was intended as a reference to Mrs Chambers banging on the wall of his unit at night time.

For some time before Mrs Chambers' stabbing, the accused had been hearing loud knocking or banging noises in his unit. The noises occurred at night. The accused believed that they were made by Mrs Chambers. He had spoken to several people about the noises and told those people that the noises had become so bad that he was planning to move to Mount Isa. Indeed, he had sold most of his furniture and bought a bus ticket before the stabbing took place.

A plumber testified that suddenly turning off a tap in the accused's unit caused a knocking noise which is known as "water hammer". The sudden turning off of the tap caused a wave to travel through the water in the pipe with the result that the washer in the tap reverberated noisily. The plumber gave evidence to the effect that it was possible that there was "water hammer" in the accused's unit which was caused by the sprinkler system installed in the grounds of the block of units where the accused lived. The sprinkler system had six sprinkler stations. It was programmed so that at 2 am one sprinkler station would come on, remain on for thirty minutes, and then be shut off suddenly by an electrical solenoid at the same time as the next sprinkler station would come on. This cycle would be repeated for all six sprinkler stations. This meant that there was a station being suddenly shut off once every thirty minutes from 2.30 am until 5 am.

At his trial, the accused did not deny that he had stabbed the deceased. However, he sought a conviction for manslaughter on the ground of diminished responsibility. Section 37 of the Criminal Code NT provides to the effect that a person who suffers from diminished responsibility shall not be guilty of murder and is guilty of manslaughter only. In order for the accused to make out the defence of diminished responsibility, it was necessary for him to prove that at the time of the stabbing he was "in such a state of abnormality of mind as substantially to impair his capacity to understand what he was doing or his capacity to control his actions or his capacity to know that he ought not do the act" [2] .

In support of the accused's claim that he had diminished responsibility at the time of the stabbing, three experts (Dr Vine, Dr Walton and Mr Taylor) gave evidence that the accused suffered: (a) cognitive defects arising from frontal lobe damage which was the result of alcohol and benzodiazepine abuse; (b) clinical depression; and (c) a delusional disorder that the deceased was persecuting him by deliberately banging on the walls of her unit at night.

Medical witnesses for the Crown were not given the opportunity to examine the accused before trial and were not provided with copies of the reports of the defence witnesses prior to the trial. The trial judge rejected an application by the Crown that it be permitted to adduce its medical evidence in reply to the medical evidence for the accused. As a result, the Crown was compelled to lead its medical evidence as part of its case in chief, even though the accused bore the onus of establishing diminished responsibility. The learned trial judge erred in refusing to permit the Crown to lead evidence in reply to the evidence of the defence experts. A plea of diminished responsibility, like a plea of insanity, is a plea of confession and avoidance. Any person, relying on the plea, must prove it. In this Court, the accused conceded that this was so. Until an accused person tenders evidence in support of the claim of diminished responsibility, the Crown has no issue to meet. It is not like the common law "defences" of provocation or "self-defence" which the Crown must negative once they are fairly raised on the evidence.

In substance, the Crown medical evidence denied that the accused had frontal lobe damage, denied that he had a delusional disorder, and denied that he suffered from depression to any relevant extent. The Crown witnesses considered that there were rational external explanations for the noises which the accused had said he had been hearing (such as the water hammer in the pipes). Accordingly, he was not delusional. While the accused may have been unhappy and angry at the time of the stabbing because of his general life situation and the fact that he mistakenly considered that the noises were being made by the deceased, he did not fall within s 37 of the Code. This was so even though the accused was intoxicated and perhaps under the influence of benzodiazepines at the time of the stabbing.

The Crown also relied on evidence that two "CT" scans carried out on the accused had not revealed any physical abnormality in his brain, in the frontal lobes or elsewhere.

Facts upon which diagnoses of defence witnesses were based

The diagnoses of the defence witnesses were substantially based upon out-of-court statements made to them by the accused, who did not testify at the trial. In this Court, counsel for the accused pointed to six categories of out-of-court assertions and answers made by the accused which formed a substantial part of the factual basis for the opinions of the defence expert witnesses. Those categories were: (a) assertions by the accused that he had no memory of the stabbing; (b) assertions by the accused that he believed that the deceased was banging on the walls of her residence in order to upset him and make him leave his flat; (c) assertions by the accused regarding his alcoholism; (d) assertions by the accused that he had been using large quantities of benzodiazepines; (e) assertions by the accused of a history of insomnia, poor appetite, social withdrawal and despondency; (f) answers given to a "Lezak" test (which is a test designed to measure the genuineness of the accused's responses to other neuropsychological clinical tests).

Because the accused did not give evidence at trial, these assertions and answers were proved by the medical, police and other witnesses to whom they were made. Such evidence is hearsay. At common law, much of it would be inadmissible as self-serving out-of-court statements. However, the Crown did not object to the assertions and answers being proved by out-of-court statements although, of course, the truth or falsity of the assertions and some answers were in issue. If the jury were satisfied that a substantial part of this material was false, the factual basis for the diagnoses of the defence expert witnesses was significantly undermined. The accused bore the onus of proof on the issue of diminished responsibility. The jury could not rationally accept the opinions of his expert witnesses unless they first accepted the factual basis for the opinions of those witnesses. If the jury rejected the substance of his assertions and answers, there was nothing to support the diagnoses of the defence witnesses except those statements made by the accused immediately or shortly after the stabbing and his answers to the investigating police officers that were tendered as admissions against him.

Whether the jury accepted the truth of the answers and assertions relied on by the expert witnesses was largely, if not wholly, dependent on the view that they formed about the credibility of the accused. His credibility was therefore an issue of great importance at the trial. Because that is so, the accused contends that his trial miscarried when the learned trial judge failed to direct the jury that they could use evidence of his good character to conclude that his out-of-court assertions were credible.

In support of his claim that he was a person of good character, the accused adduced evidence that he had no previous convictions for a criminal offence other than a conviction for drink-driving in 1975, and evidence that he was not "adversely known to the police". He also adduced evidence of his character and personality from those who knew him. It will later be necessary to refer to this character evidence in more detail.

The trial judge's direction to the jury on good character

In this Court, the argument of counsel for the accused assumed that counsel at the trial had asked her Honour to direct the jury that the evidence of good character was relevant to both: (a) the improbability of the accused having committed the instant offence, or, as counsel for the accused put it, the improbability that the accused is a person "who will make a deliberate choice to kill in a rational state"; (b) assessing the credibility of the accused in making the assertions and answers which were the basis of the opinions of his expert witnesses.

However, the remarks of counsel in seeking directions and in his closing address suggest that he placed little reliance on the character evidence to support the credibility of the answers and assertions that were the basis of the expert opinions. In the course of his submissions for re-directions, counsel said:

[I]t would be my submission that the aspect of the improbability of committing the instant offence, having a history of good character for sixty-odd years, is of considerable significance.

The aspect of his credibility is probably of lesser significance, having regard to the nature of the interview itself that has been severely criticised by me as showing a lack of credibility, but, in any event, it is the primary aspect of the evidence that should be brought to the jury's attention and one which as a matter of law, he is entitled to.

I thought the most convenient repository of the law in relation to this is a decision of R v Murphy [3]

The situation in Murphy's case, obviously, was that it's the aspect of credibility which was the most significant. In my submission, in this case it's the reverse; it's the aspect of probability or inherent probability of the commission of the offence. (Emphasis added.)

In reply to counsel for the Crown, counsel for the accused said:

[T]he aspect of credibility, while I say it [is] not the primary consideration, nonetheless, is also significant, in terms of considering his explanation, not just to Mr Newman [4] but to the others. 1. Mr Newman was the detective in charge of the investigation.

The learned trial judge adjourned for a short time after hearing those submissions. On her return to court, her Honour said that she would "give the direction as sought by the defence". However, her Honour failed to give a direction as to how the jury might use the character evidence in assessing the accused's credibility. Her Honour said:

[W]hen you consider that evidence as to good character you are entitled to consider the improbability of Roy Melbourne committing the instant offence, having a history of good character of some sixty-one years, and that this is of considerable significance.

Counsel for the accused did not seek any further direction on credibility. For the purpose of the appeal to the Court of Criminal Appeal, junior and senior defence counsel swore affidavits in which they said that at the trial they overlooked the fact that the direction which her Honour gave differed from that which they had sought and which Thomas J had indicated that she would give. Rule 86.08 of the Supreme Court Rules NT provides that:

No direction, omission to direct or decision in relation to the admission or rejection of evidence of the Judge of the court of trial shall, without the leave of the Court of Criminal Appeal, be allowed as a ground for appeal, or for an application for leave to appeal, unless objection was taken at the trial to the direction, omission or decision by the party appealing or applying for leave to appeal. Pursuant to this rule, the Court of Criminal Appeal gave leave to the accused to object to her Honour's failure to direct the jury that the character evidence was relevant in assessing the accused's credibility.

There is no reason to doubt the sworn statements of counsel that they did not seek any tactical advantage in not pressing for a further direction on character and that they did not intend to abandon the credibility aspect of the character evidence. Moreover, while counsel conceded that the aspect of credibility was "probably of lesser significance", his request for a direction on character included a request that the judge direct the jury on the "credibility" as well as the "probability" significance of the character evidence. No doubt it seems likely that, in seeking a credibility direction, counsel did not have in mind the answers and assertions to the expert witnesses. His concern seems to have been directed to the accused's explanation for his conduct at or shortly after the killing rather than the histories given to the doctors or the accused's answers to the Lezak test. But the accused's explanations for his conduct were not without significance for the medical opinions. Moreover, if the jury thought that his explanations were credible, they may well have accepted the accused's assertions and answers given to the expert witnesses. That being so, the issue is whether the trial judge erred by failing to direct the jury that the accused's character was relevant in assessing his credibility and, if so, whether, in all the circumstances of the case, there has been a miscarriage of justice.

Hitherto, Australian trial judges have had a discretion as to the directions that they should give to the jury concerning the use to be made of good character evidence. In Simic v The Queen [5] , this Court held that no miscarriage of justice had occurred when the trial judge had failed to direct the jury as to the manner in which they could use evidence that the accused was a person of good character. In a joint judgment, Gibbs, Stephen, Mason, Murphy and Wilson JJ said [6] :

There is no rule of law that in every case in which evidence of good character is given the judge must give a direction as to the manner in which it can be used No doubt, speaking generally, it is right to add that if such a direction is asked for it would be wise to give it.

In the present case no direction as to the evidence of the applicant's good character was asked for. There is no reason to believe that the jury would not have understood that a man of good character would be unlikely to commit a crime of savage violence such as that with which the applicant was charged. In other words, there is no reason to conclude that the jury would have failed to give the evidence as to good character such weight as it deserved.

1. (1980) 144 CLR 319. 2. Simic (1980) 144 CLR 319 at 333-334.

23. Counsel for the accused has submitted that this Court should no longer regard the giving of directions as to character as a matter for the discretionary judgment of the trial judge. Instead, counsel submitted that the Court should follow the appellate courts in England and New Zealand and hold that there is now a rule of practice that such a direction should be given.

Directions on good character evidence in the United Kingdom and New Zealand

24. In R v Berrada [7] , the English Court of Appeal held that, in cases where the accused has adduced evidence of good character, the trial judge must direct the jury how to use that evidence. The Court held that, where the credibility of the accused was in issue, the jury must be given a direction about the relevance of the accused's previous good character to his or her credibility [8] . Furthermore, it was proper but not obligatory "for the judge to refer to the fact that the previous good character of the [accused] might be thought by them to be one relevant factor when they were considering whether he was the kind of man who was likely to have behaved in the way that the prosecution alleged" [9] .

1. (1989) 91 Cr App R 131. 2. R v Berrada (1989) 91 Cr App R 131 at 134. 3. R v Berrada (1989) 91 Cr App R 131 at 134.

25. In R v Vye [10] , the Court of Appeal expanded the rule in Berrada by holding that the credibility limb of the direction applied in assessing the reliability of exculpatory statements made to the police even if the defendant did not testify at his or her trial.

26. But what constitutes "good character" for the purpose of these directions? In R v Aziz [11] , Lord Steyn (with whom the other members of the House of Lords agreed) described [12] the equation of a lack of a criminal record with evidence of good character as the "usual case". In Aziz, three persons stood trial for fraudulently evading payment of VAT. None of them had prior convictions. But one of them admitted to having knowingly made a false mortgage application and to having lied to customs officers during an interview; another acknowledged that he had made false income tax returns and had submitted a false mortgage application form. The House of Lords held that, where the accused has no prior convictions, the trial judge is prima facie bound to direct the jury in respect of the accused's good character [13] . However, in order to avoid the absurdities that could arise from an absolute rule, their Lordships held [14] that the trial judge has "a residual discretion to decline to give any character directions in the case of a defendant without previous convictions if the judge considers it an insult to common sense to give directions in accordance with Vye. "

1. [1996] AC 41. 2. R v Aziz [1996] AC 41 at 51. 3. R v Aziz [1996] AC 41 at 53. 4. R v Aziz [1996] AC 41 at 53.

27. In his article, "What Constitutes a Good Character?", Mr Roderick Munday [15] points out some of the anomalies which have arisen from equating good character with an absence of convictions. For example, in R v Anderson [16] , a policeman admitted picking up a woman in his police car while he was on duty and having intercourse with her in the car. He was charged with, but denied, raping her. He claimed that the intercourse was consensual. Because the policeman had no previous convictions, the Criminal Division of the Court of Appeal held that he was entitled to a direction that he was of good character. As Mr Munday points out, "a police officer who admitted to conduct unbecoming was none the less held in law to be entitled to be treated as someone of good character" [17] .

1. Munday, "What Constitutes a Good Character?" [1997] Criminal Law Review 247. 2. [1990] Crim LR 862 (note). 3. Munday, "What Constitutes a Good Character?" [1997] Criminal Law Review 247, at p 250 .

In R v Falealili [18] , a majority of the New Zealand Court of Appeal held [19] that when evidence of good character was adduced:

an appropriate direction should be given as to its use. Generally that will cover both limbs of credibility and propensity. No particular form of words is necessary, and because of the variety in the circumstances in which the need will arise, the direction will no doubt be tailored to meet those circumstances. 1. [1996] 3 NZLR 664. 2. R v Falealili [1996] 3 NZLR 664 at 667.

In Falealili , the Court of Appeal did not accept that an absence of criminal convictions was synonymous with "good character". The majority said [20] :

We think there are logical difficulties with the proposition that an absence of previous convictions is in itself evidence establishing a person's good character. It may be a factor in assessing good character, but standing on its own it is generally neutral. A person of bad repute may well have no convictions. We do not think it necessary for directions to be given merely because absence of previous convictions has been elicited. The need will arise where evidence relating to character has been adduced which, if accepted by the jury, could properly be relevant or probative in determining whether guilt has been proved. That after all is the basis of its admissibility. 1. R v Falealili [1996] 3 NZLR 664 at 667.

In my opinion, notwithstanding the rules laid down in these English and New Zealand cases, this Court should not depart from the rule that a judge is not obliged to direct the jury concerning the accused's good character. The preferable position is that the trial judge must retain a discretion as to whether to direct the jury on evidence of good character after evaluating its probative significance in relation to both (a) the accused's propensity to commit the crime charged; and (b) the accused's credibility.

The judge may conclude that the good character evidence adduced is of probative significance in relation to (a) only, (b) only, both (a) and (b) or neither (a) nor (b), and can direct (or not direct) the jury accordingly. Whether the discretion has miscarried in a particular case will depend upon the facts of that case. But Australian courts should not now introduce a rule that a direction on character is always required once the accused has adduced evidence of good character.

Two considerations lead me to this conclusion. First, the difference between the use of good character evidence and the use of bad character evidence in a criminal trial is logically anomalous and, while that difference is too deeply rooted in the law to be removed by judicial decision, it should not be widened. Second, in cases where good character evidence has no logical connection with the elements of the offence, a mandatory direction is likely to divert the jury from properly evaluating evidence which more directly and logically bears upon the guilt of the accused and, in cases like R v Anderson [21] and R v Aziz [22] , such a direction may even confuse the jury.

1. [1990] Crim LR 862 (note). 2. [1996] AC 41.

33. In its strict sense, character refers to the inherent moral qualities of a person or what the New Zealand Law Commission has called " disposition - which is something more intrinsic to the individual in question" [23] . It is to be contrasted with reputation, which refers to the public estimation or repute of a person, irrespective of the inherent moral qualities of that person [24] . As the last of the above passages from R v Falealili [25] demonstrates, however, the common law courts have not always drawn a distinction between character and reputation in a criminal context [26] . The confusion can be traced to R v Rowton [27] where a majority of the Full Court of the Crown Cases Reserved held that in a criminal trial the evidence for or against a person's good character must be confined to his or her general reputation. This is the established rule although, as this Court pointed out in Attwood v The Queen [28] , the limitations inherent in the rule are not observed in practice. In New South Wales, the legislature long ago reversed the common law rule [29] .

1. Preliminary Paper 27, Evidence Law: Character and Credibility (1997), par 99 (emphasis in original). 2. Plato Films Ltd v Speidel [1961] AC 1090 at 1138, per Lord Denning. 3. [1996] 3 NZLR 664. 4. In the field of defamation, however, the distinction is well recognised (see, eg, Plato Films Ltd v Speidel [1961] AC 1090). 5. (1865) Le & Ca 520 [ 169 ER 1497]. 6. (1960) 102 CLR 353 at 359. 7. In 1900, s 413 of the Crimes Act 1900 NSW relevantly provided: "Every witness examined as to character may give evidence not only as to the general repute of such person, but also as to the witness's own knowledge of his habits, disposition, and conduct." The matter is now dealt with in s 110(1) of the Evidence Act 1995 NSW which allows a witness to give evidence as to his or her own opinion of the accused's character.

34. In the criminal field, the common law has also tended to treat people as one-dimensional personalities who have either good or bad characters or dispositions. This tendency has been checked in the field of defamation, where the issue is reputation and not character and where the plaintiff obtains damages for the injury to reputation in the particular sector of the plaintiff's life to which the libel refers [30] . But the tendency continues to prevail in the criminal law, where a person is regarded as having either a good character or a bad character. In the absence of a statutory provision to the contrary [31] , for example, a convicted sex-offender will be treated as a person of bad character in a trial for embezzlement although there is overwhelming evidence of that person's honesty. Conversely, in England a person without convictions may be entitled to a good character direction although his or her general conduct suggests the contrary.

1. Plato Films Ltd v Speidel [1961] AC 1090. 2. In trials where the Evidence Act 1995 Cth and NSW applies, s 110(3) enables the accused to adduce evidence that he or she "is a person of good character in a particular respect".

35. What Lord Radcliffe said in Plato Films Ltd v Speidel [32] concerning evidence of general reputation in defamation cases seems equally applicable to evidence of good and bad character in criminal cases. His Lordship said:

The difficulty is that "general evidence of reputation" does not convey an idea of any content. Life not being a morality play or a Victorian melodrama, men do not enjoy reputations for being bad or good simpliciter: nor if they did, would the proof of such generalities throw any light upon the loss of reputation suffered from a particular libel. Similarly, in many criminal cases, evidence that a person is of good character in the general sense recognised by the common law throws little, if any, light upon the probability whether he or she committed the crime in question. 1. [1961] AC 1090 at 1130.

Treatment of evidence of bad character or criminal propensity

36. For more than a century, the common law has drawn a distinction between the admissibility of evidence of good character and the admissibility of evidence of bad character in a criminal trial. Evidence of good character is readily admitted because it is regarded as tending to prove that the accused is unlikely to have committed the crime in question. Evidence of bad character is admitted only in exceptional circumstances even where the courts regard it as tending to prove that the accused is likely to have committed the crime in question.

37. The common law has developed strict rules for the admissibility of evidence designed to prove that, by reason of his or her character or propensities, the accused is likely to have committed the crime with which he or she is charged. In Makin v Attorney-General (NSW) [33] , Lord Herschell said that the prosecution cannot:

adduce evidence tending to shew that the accused has been guilty of criminal acts other than those covered by the indictment, for the purpose of leading to the conclusion that the accused is a person likely from his criminal conduct or character to have committed the offence for which he is being tried. On the other hand, the mere fact that the evidence adduced tends to shew the commission of other crimes does not render it inadmissible if it be relevant to an issue before the jury, and it may be so relevant if it bears upon the question whether the acts alleged to constitute the crime charged in the indictment were designed or accidental, or to rebut a defence which would otherwise be open to the accused. 1. [1894] AC 57 at 65.

38. In Maxwell v Director of Public Prosecutions [34] , this statement was said to give effect to "one of the most deeply rooted and jealously guarded principles of our criminal law". In this Court, its status as a fundamental principle has been confirmed in numerous cases [35] .

1. [1935] AC 309 at 317. 2. See, eg, Burrows v The King (1937) 58 CLR 249 at 253; Perry v The Queen (1982) 150 CLR 580 at 585; Pfennig v The Queen (1995) 182 CLR 461 at 475.

As the passage from Makin v Attorney-General (NSW) [36] demonstrates, evidence disclosing the bad character of the accused is sometimes admissible. However, courts, including this Court, have consistently held that evidence of the bad character of the accused or the propensity of the accused to commit criminal acts is only admissible if strict conditions are fulfilled. In Pfennig v The Queen [37] , Mason CJ, Deane and Dawson JJ held that propensity evidence is admissible only if it possesses a particular probative value or cogency such that, if accepted, it bears no reasonable explanation other than the inculpation of the accused in the offence charged. In Hoch v The Queen [38] , the Court said that the probative force of such evidence when admitted:

lies in the fact that the evidence reveals "striking similarities", "unusual features", "underlying unity", "system" or "pattern" such that it raises, as a matter of common sense and experience, the objective improbability of some event having occurred other than as alleged by the prosecution. This statement was approved by the majority in Pfennig [39] . 1. [1894] AC 57. 2. (1995) 182 CLR 461 at 475. 3. (1988) 165 CLR 292 at 294-295. 4. (1995) 182 CLR 461 at 482.

The "no other reasonable explanation" test requires the judge to come to a view as to the guilt of the accused before the evidence in question is admitted. If, at the stage of determining whether the evidence is admissible, the judge decides that there is no "reasonable explanation" for the evidence other than inculpation of the accused, the evidence will be admitted. The judge has then, in effect, determined that the accused is guilty of the charges although, of course, it is for the jury to determine the ultimate question of the guilt or innocence of the accused on the whole of the evidence. Where the trial is by a judge without a jury, he or she must also examine the whole of the evidence before finding the accused guilty, notwithstanding that he or she has already decided that there is no reasonable explanation for the disputed evidence other than the accused's guilt.

In Perry v The Queen [40] , Gibbs CJ explained the rationale for the rule excluding evidence of bad character or propensity as follows:

Evidence that an accused person has a propensity to commit crimes of the sort with which he is charged, or is the sort of person who is likely to commit such crimes, would ordinarily be regarded as relevant to the question whether he did commit the offence in question. Such evidence is excluded, not because it is irrelevant, but because it is likely to be unfairly prejudicial to the accused. 1. (1980) 150 CLR 580 at 585.

The rationale of the common law rule has greatly influenced the approach of courts to statutory provisions that enable bad character evidence to be tendered in evidence. Dawson v The Queen [41] , where this Court considered s 399(e) of the Crimes Act 1958 Vict, is a good example. That sub-section relevantly stated that an accused person appearing as a witness:

shall not be asked, and if asked shall not be required to answer, any question tending to show that he has committed or been convicted of or been charged with any offence other than that wherewith he is then charged, or is of bad character, unless -

(ii) he has personally or by his advocate asked questions of the witnesses for the prosecution with a view to establish his own good character, or has given evidence of his good character, or the nature or conduct of the defence is such as to involve imputations on the character of the prosecutor or the witnesses for the prosecution.

In Dawson, the prosecution relied on a verbal admission by the accused that a record of interview accurately recorded admissions that he was involved in a breaking and entering offence. One issue at the trial was whether, by denying making the admission, the accused had rendered himself liable to cross-examination on his past convictions and bad character under s 399. Dixon CJ concluded that the section did not cover "inferences, logical implications or consequential deductions which may spell imputations against the character of witnesses" [42] , and that imputations on the prosecution witnesses must be "an element or ingredient in the defence or what arises from the manner in which the defence is conducted" [43] for the accused to be exposed to cross-examination on past convictions and character. In support of his conclusion, Dixon CJ said:

It is the thesis of English law that the ingredients of a crime are to be proved by direct or circumstantial evidence of the events, that is to say, the parts and details of the transaction amounting to the crime, and are not inferred from the character and tendencies of the accused. [44] 1. Dawson (1961) 106 CLR 1 at 9-10. 2. Dawson (1961) 106 CLR 1 at 9. 3. Dawson (1961) 106 CLR 1 at 16.

The explanation by Gibbs CJ of the rationale of the rule concerning bad character evidence demonstrates that, in determining the admissibility of such evidence, the courts have accepted, consciously or unconsciously, that character evidence is likely to divert the tribunal of fact from the true issues in the case. It is likely to divert the jury from properly evaluating the strength or weakness of evidence that more directly bears on whether or not the accused committed the crime in question.

In the field of similar fact evidence, Mr Rajiv Nair has drawn a distinction between the "moral" and "descriptive" elements of such evidence [45] . He contends that similar fact evidence contains a "descriptive" element in the sense that it describes a feature associated with an accused person or his circumstances which exists independently of the circumstances surrounding the commission of the alleged offence and which is relevant to the offence charged. It also contains a "moral" element in the sense that it carries connotations as to the moral desirability or otherwise of possessing certain qualities. In Mr Nair's view, the "prejudicial effect" of similar fact evidence arises primarily out of its moral quality while the probative force arises out of its descriptive quality.

1. Nair, "Weighing Similar Fact and Avoiding Prejudice", Law Quarterly Review, vol 112 (1996) 262, at p 273 .

46. The opportunity for prejudice arising out of this moral element of character evidence was reduced by the course taken in Pfennig to apply a restrictive test for the admissibility of similar fact evidence and the course taken in Dawson to read the statutory exception narrowly. In both cases, the Court was concerned to ensure that evidence upon which a moral judgment concerning the accused can be made is not admitted unless it is also rationally connected with "the parts and details of the transaction amounting to the crime" [46] . Unless such a connection exists, the moral element of the character evidence is likely to be used to make an irrational connection between the person's character and his or her guilt. If the descriptive element of the character evidence shows an irresistible rational connection with the "parts and details of the transaction amounting to the crime" [47] , however, bad character or propensity evidence will be admitted. Its admissibility can be justified because, where the rational connection is sufficiently strong, the moral element of such evidence is unlikely to divert the jury or other tribunal of fact from its proper function.

1. Dawson v The Queen (1961) 106 CLR 1 at 16. 2. Dawson v The Queen (1961) 106 CLR 1 at 16.

47. Given the common law's acceptance of the diversionary effect of bad character evidence, the manner in which it allows good character evidence to be used in a criminal trial is anomalous. Good character evidence is not subject to the stringent evaluation of its probative force that is applied to evidence of bad character. It is admitted condition free. Yet there is no logical or legal reason for drawing a distinction between the conditions for admitting bad character evidence and the conditions for admitting good character evidence. Furthermore, as Kirby J points out in his judgment, empirical psychological studies now deny that character is as accurate a predictive tool as earlier generations so confidently believed [48] . The unconditional right of an accused person to tender good character evidence must be regarded as an indulgence granted to the accused which continues to be maintained for historical reasons. The basis of the rule for admitting evidence of good character is not logic but the "policy and humanity" [49] of the common law.

1. Even as great an authority on the law of evidence as Professor Wigmore had no doubt that evidence of character or disposition pointed to the probability of the accused having committed the offence charged. (Evidence in Trials at Common Law, Tillers rev (1983), vol 1a, s55.) 2. R v Rowton (1865) Le & Ca 520 at 541 [ 169 ER 1497 at 1506], per Willes J (dissenting) cited in Attwood v The Queen (1960) 102 CLR 353 at 359.

It would be anomalous if before evidence of bad character or criminal propensity is even admitted it is subject to a rigorous evaluation of its probative significance, and yet good character evidence of dubious probative value is not only admitted, but is required to be the subject of a mandatory direction favourable to the accused even if the trial judge considers that the direction is not warranted in the circumstances of the case.

In my opinion, the distinction between "moral" elements and "descriptive" elements formulated by Mr Nair can be generalised from similar fact evidence to character evidence in general. That being so, if the law of evidence was a logically coherent body of doctrine, good character evidence would not be admitted unless as a minimum it tended to negative some part or detail "of the transaction amounting to the crime" [50] . But it is too late in the day to hold that good character testimony must meet such conditions to be admissible. That does not mean, however, that in defiance of logic and modern psychological opinion it should automatically be treated as if it did negative the parts or details of the transaction.

The dissent of Thomas J in R v Falealili [51] recognises the necessity for good character evidence to have probative value before the judge should give the jury a direction as to the manner of using it. His Honour said:

Consequently, if the evidence of the accused's good character is both probative and relevant the Judge will, almost as a matter of course, direct the jury as to its significance in summing up the defence case. It would be unfair not to do so. If, on the other hand, the purported character evidence is lacking in probative force and of remote relevance to the charge in issue, the Judge may decide that a good character direction is not warranted. Or the Judge may consider that it would be prudent to proffer a good character direction, but then to qualify it in order to put it in perspective having regard to the circumstances of the case. To proscribe that, whenever character evidence is adduced or elicited, a good character direction should be given and that it must generally embrace both the credibility and propensity limbs of the direction is an unnecessary fetter on that discretion. 1. [1996] 3 NZLR 664 at 671-672.

The majority judges in Falealili also recognised the importance of the probative effect of good character evidence. Although in that case their Honours stated [52] that a character direction "should be given", they also said that the need for a direction "will arise where evidence relating to character has been adduced which, if accepted by the jury, could properly be relevant or probative in determining whether guilt has been proved." In doing so, the Court has recognised the importance of the probative value of the evidence. Consequently, although in form the New Zealand Court of Appeal appears to have adopted a mandatory rule, its formulation allows the trial judge to determine when the threshold of "good character" has been met, and thus when the mandatory direction is required, by reference to the probative value of the evidence.

There are other advantages in not having a mandatory direction. It avoids the need to attempt to define in advance what is "good character" and thus the circumstances in which the mandatory directions will be invoked. Defining the absence of a criminal record as equivalent to good character required the House of Lords in R v Aziz [53] to create a "residual discretion" to avoid the absurdities which resulted from such a definition.

Application of these principles to the accused's case

In my opinion, the character evidence relating to Mr Melbourne was not of such probative significance in relation to his credibility as to require the trial judge to give a direction that the evidence bore favourably upon Mr Melbourne's credibility. The evidence was that the accused had no previous convictions other than a conviction for drink-driving in 1975 and was not "adversely known to the police". Various descriptions of his character and personality were given by those who knew him, such as: (a) evidence from Mr Gooch that the accused was a "quiet man", a man who was "always gentle", and who, apart from this occasion, had "never" been "aggressive"; (b) evidence from Mrs Barnes that the accused was "very quiet"; (c) evidence from Mr Daniels that the accused was "a very amiable sort of person"; (d) evidence from Mrs Hinde that the accused was "a very quiet, well-behaved gentleman".

None of this evidence had any direct probative bearing on the truthfulness or credibility of the accused. It was all directed to the unlikelihood that he would commit the offence charged. The trial judge gave an adequate direction in this regard. Whether or not the trial judge intended, but forgot, to give a credibility direction with respect to the character evidence, no miscarriage of justice has occurred. If her Honour had given such a direction, it would have given the accused an advantage to which in point of law he was not entitled. Not only was this not a case requiring a credibility direction, in my opinion it would have been a wrongful exercise of discretion to have given it.

On 13 June 1996, the appellant was convicted on a charge of murder, as provided by s 162 of the Criminal Code NT (the Code) after a trial before a judge and jury in the Supreme Court of the Northern Territory. The provisions of the Evidence Act 1995 Cth did not apply at the trial [54] .

Evidence was given which was treated as going to the good character of the appellant. In directing the jury with respect to that matter, the trial judge said:

The first matter is a direction I propose to give you in respect of what use you make of the evidence you heard during the course of these proceedings, of the good character of the accused, Roy Melbourne.

I do not propose to go back and remind you again of that evidence, but there has been evidence of course from witnesses relating to his not being a violent or aggressive person, and references to the fact that he was a quiet and amiable man.

I will specifically remind you of the evidence that was given by Sergeant Newman. Sergeant Newman gave evidence to the effect that he had, after Mr Melbourne was arrested, done a check of his record, his criminal record, and that there were no matters on that record other than a conviction for an offence of exceed .08 with a reading of .23 in 1975. Sergeant Newman had gone on to say that from his investigations there were no other convictions for any other matters in the Northern Territory or anywhere, and that Mr Melbourne was not adversely known to police.

The direction I am giving you is this: that when you consider that evidence as to good character, that you are entitled to consider the improbability of Roy Melbourne committing the instant offence, having a history of good character of some sixty-one years, and that this is of considerable significance.

No fuller direction was sought at the trial. Nevertheless, in this Court, as in the Northern Territory Court of Criminal Appeal, the appellant submits that the trial judge erred in her direction with respect to his "good character".

The appeal should be dismissed. I agree with the reasons for judgment of Hayne J and would add only the following.

As Hayne J points out, the appellant did not deny that he inflicted the injuries causing the death of his neighbour. The "previous good character" of the appellant in this Court was said to go to the issue whether the jury should accept that he had sought to tell the truth when interviewed by police and later by experts subsequently called to give evidence of his mental condition. The appellant put forward a defence of diminished responsibility within the meaning of s 37 of the Code [55] . It may be accepted that if, in these conversations, the appellant had been trying to establish some false basis for a later plea of diminished responsibility, that circumstance may have had some significance, as a matter of inference, in deciding his mental condition at the time he slew the victim.

1. This states: "When a person who has unlawfully killed another under circumstances that, but for this section, would have constituted murder, was at the time of doing the act or making the omission that caused death, in such a state of abnormality of mind as substantially to impair his capacity to understand what he was doing or his capacity to control his actions or his capacity to know that he ought not do the act, make the omission or cause that event, he is excused from criminal responsibility for murder and is guilty of manslaughter only."

61. In such a context, what is meant by an assertion that the appellant is a person of "previous good character"?

A defendant's character, then, as indicating the probability of his doing or not doing the act charged, is essentially relevant. In point of human nature in daily experience, this is not to be doubted. The character or disposition - ie, a fixed trait or the sum of traits - of the persons we deal with is in daily life always more or less considered by us in estimating the probability of their future conduct. In point of legal theory and practice, the case is no different. 1. Evidence in Trials at Common Law, Tillers rev (1983), vol 1a, s55 (footnote omitted).

63. Nevertheless, to those not versed in the ways of the common law, it may appear curious that legal consequences follow from the attachment to a designated individual, and without further analysis, of the description "good character" or "bad character". First, this appears to assume polarities with no space for occupation by those whose frailties place them somewhere towards the centre of a continuum. Secondly, it allows too little scope for the infinite variety of mental processes which lead to action or inaction, and assumes that people act across a range of circumstances in conformity with a measurable trait which can be the subject of testimony. Thirdly, in the development of the English language, and thus of the common law, the term "character" has had various shades of meaning. The Oxford English Dictionary [57] gives eleven uses of the term in a figurative sense in addition to its primary and literal senses of a distinctive mark or symbol. In particular, in its figurative sense, "character" may identify (i) a trait which serves as an index to the essential or intrinsic nature of an individual, (ii) the sum of such traits, or (iii) the estimate put upon an individual as a matter of repute.

64. In the law, the notion of "character" takes varying significance and shades of meaning from particular fields of discourse and the particular fact in issue. It may be said that "character", that which marks out an individual, may not correspond with the reputation attributed to that person. However, as will appear, the law does not always clearly distinguish between the two, nor indicate the probative force to be attributed to whichever of them is to be established as a fact in issue, nor specify the evidentiary means, including permissible inference, by which that fact in issue may be proved.

65. The matter is well put by the New Zealand Law Commission in its Preliminary Paper, Evidence Law: Character and Credibility [58] . Paragraphs 99 and 100 include the following:

On the one hand, the law distinguishes between evidence of general reputation and evidence of individual opinion and, in the case of the defendant in criminal proceedings, has historically recognised only the former [59] . On the other hand, it is not always clear what is meant by reputation. On occasion, it appears to be used interchangeably with character. It may be important therefore to distinguish between character as public estimation - which is perhaps more correctly referred to as reputation - and character as disposition - which is something more intrinsic to the individual in question [60] .

In actions for defamation the first meaning is paramount, since it is the public perception of an individual which the law of defamation protects. The second meaning is of primary significance when a party seeks to offer similar fact evidence to show an individual's propensity to commit certain offences In both cases, the evidence of reputation goes to the issue. But reputation has also traditionally been a factor indicative of a person's truthfulness. Its meaning in this context seems to be an amalgam of public estimation and individual disposition. (Emphasis in original.)

1. Preliminary Paper 27 (1997). 2. R v Rowton (1865) Le & Ca 520 [ 169 ER 1497]. 3. See Plato Films Ltd v Speidel [1961] AC 1090 at 1128, 1138.

The issue in a proceeding may be whether an individual has the good character required for admission to pursue a particular profession or calling. Here the concern is not with disposition to perform particular acts with a requisite intention. Nor is the question simply one of the opinion others may have of the individual in question. In Ex parte Tziniolis; Re Medical Practitioners Act [61] , Holmes JA said:

The Act provides for the circumstances in which the name of a registered medical practitioner may be removed from the register and the expression "infamous conduct in a professional respect" has been used to define such conduct. "Good character" is not a summation of acts alone, but relates rather to the quality of a person. The quality is to be judged by acts and motives, that is to say, behaviour and the mental and emotional situations accompanying that behaviour. However, character cannot always be estimated by one act or one class of act. As much about a person as is known will form the evidence from which the inference of good character or not of good character is drawn. His Honour emphasised that the court was not there dealing with "good character" in some particular sense developed by the criminal law or by the law of defamation [62] . 1. (1966) 67 SR (NSW) 448 at 475-476. 2. Ex parte Tziniolis (1966) 67 SR (NSW) 448 at 475.

With respect to the latter, in Plato Films Ltd v Speidel [63] , Lord Radcliffe perceived the issue as being whether a defendant may offer in mitigation of damages evidence which bears upon the disposition of the plaintiff, as distinct from his reputation, or only such evidence as bears upon his reputation [64] . His Lordship concluded that the defendant is confined to the latter species of evidence. However, in the course of his speech, Lord Radcliffe observed [65] :

The difficulty is that "general evidence of reputation" does not convey an idea of any content. Life not being a morality play or a Victorian melodrama, men do not enjoy reputations for being bad or good simpliciter: nor if they did, would the proof of such generalities throw any light upon the loss of reputation suffered from a particular libel. Lord Denning referred to the distinction in meaning between "character" and "reputation", saying [66] : A man's "character", it is sometimes said, is what he in fact is, whereas his "reputation" is what other people think he is. If this be the sense in which you are using the words, then a libel action is concerned only with a man's reputation, that is, with what people think of him: and it is for damage to his reputation, that is, to his esteem in the eyes of others, that he can sue, and not for damage to his own personality or disposition. (Emphasis in original.) 1. [1961] AC 1090. 2. Plato Films [1961] AC 1090 at 1127. 3. Plato Films [1961] AC 1090 at 1130. 4. Plato Films [1961] AC 1090 at 1138.

What then of the criminal law? Statute apart, the tender of evidence as to the good reputation of the accused was permitted at a time before the accused became a competent witness [67] . In that era there could have been, in general, no question of the use by a jury of such reputation evidence in an assessment of the accused's testimonial credit [68] . Writing in this period, Starkie said [69] :

[J]uries are called upon to raise an inference in favour of a defendant in a criminal case from the goodness of his character in society; a presumption too remote to weigh against evidence which is in itself satisfactory, and which ought never to have any weight except in a doubtful case. Starkie went on to observe that the reception of such character evidence "seems to be the last remnant of compurgation" [70] , a method of trial not by jury but by wager of law whereby a sufficient number of "oath helpers" swore in favour of the character of the accused when the latter swore an oath declaring his or her innocence [71] . 1. Stirland v Director of Public Prosecutions [1944] AC 315 at 322; as to competency of parties to civil actions, see Trade Practices Commission v Abbco Iceworks Pty Ltd (1994) 52 FCR 96 at 137-139. 2. However, for example, if the prosecution tendered a confessional statement by the prisoner, it could not exclude from the tender self-serving portions; the whole became evidence for and against the prisoner: R v Higgins (1829) 3 Car & P 603 [172 ER 565]; R v Williamson [1972] 2 NSWLR 281 at 295-296. 3. Starkie, Practical Treatise of the Law of Evidence, 4th ed (1853), p 75. 4. Starkie, Practical Treatise of the Law of Evidence, 4th ed (1853), p 75. 5. Wager of law was abandoned at an early stage as a method of criminal trial in the King's courts (Thayer, A Preliminary Treatise on Evidence at the Common Law (1898), p 26) but lingered in the civil courts (as is shown by King v Williams (1824) 2 B & C 538 [107 ER 483]) until its abolition by s 13 of the Civil Procedure Act 1833 UK (3 & 4 Will 4, c 42).

This conceptual obscurity and clouded historical origin gives force to observations by Viscount Simon LC in Stirland v Director of Public Prosecutions [72] . His Lordship said [73] :

There is perhaps some vagueness in the use of the term "good character" in this connection. Does it refer to the good reputation which a man may bear in his own circle, or does it refer to the man's real disposition as distinct from what his friends and neighbours may think of him? In R v Rowton [74] , on a re-hearing before the full court [of Crown Cases Reserved], it was held by the majority that evidence for or against a prisoner's good character must be confined to the prisoner's general reputation, but Erle CJ and Willes J thought that the meaning of the phrase extended to include actual moral disposition as known to an individual witness, though no evidence could be given of concrete examples of conduct. 1. [1944] AC 315. 2. Stirland [1944] AC 315 at 324-325. 3. (1865) Le & Ca 520 [ 169 ER 1497].

In Attwood v The Queen [75] , this Court referred to Rowton and remarked that the limitations imposed by that case were probably not observed in practice. The nature of the evidence received in the present case appears to illustrate the point. No criticism is made on that count.

Indeed, in R v Ravindra [76] , the issue stemming from Rowton was faced and Gendall J held that, in addition to evidence of general reputation known to them, witnesses might give evidence as to good character based upon their personal experiences in professional, private and other dealings with the accused. His Honour observed [77] :

How is it possible for witnesses to speak of the general good reputation of an accused without simply repeating that which others say but none of whom are allowed to refer to the particular facts upon which the reputation is formed [?] The danger is that the reputation, in modern terms, is created through the repetition of myths which may have no foundation or basis upon particular facts. Without the witness being able to refer to his or her own experience so as to be able to judge the good character of an accused, the exercise becomes, with respect, illogical. 1. [1997] 3 NZLR 242. 2. R v Ravindra [1997] 3 NZLR 242 at 247-248.

In Attwood, this Court stated, with reference to the judgment of Cockburn CJ in Rowton , that evidence of good character is regarded as really bearing on the probability or improbability of guilt. Their Honours said [78] :

The expression "good character" has of course a known significance in relation to evidence upon criminal trials; for it denotes a description of evidence in disproof of guilt which an accused person may adduce. He may adduce evidence of the favourable character he bears as a fact or matter making it unlikely that he committed the crime charged. The issues in the particular case and the nature of the evidence of "good character" which is proffered will guide the process of reasoning of the tribunal of fact on the path to providing an answer to the ultimate question of whether the accused is guilty beyond reasonable doubt [79] . 1. Attwood (1960) 102 CLR 353 at 359. 2. Cross on Evidence, 5th Aust ed (1996), par 19130.

In this way, as indicated earlier in these reasons, it may be accepted that the appellant's belief in what he had told the police and the experts bore upon the mental element in the alleged offence and, in particular, upon the appellant's defence of diminished responsibility.

The trial judge instructed the jury to the effect that it was entitled to consider the improbability of the appellant having committed the offence in question, the appellant having a history of good character throughout his life and this being a matter of considerable significance. The question is whether a more extensive direction should have been given and whether failure to do so founds appealable error.

In Simic v The Queen [80] , in the joint judgment of five members of the Court, it was said that there was no rule of law that the judge must give a direction as to the manner in which good character evidence may be used in every case in which such evidence has been received. Their Honours added that they agreed with what was said on the point in R v Schmahl [81] and added [82] :

No doubt, speaking generally, it is right to add, as was said in that case, that if such a direction is asked for it would be wise to give it. 1. (1980) 144 CLR 319 at 333. 2. [1965] VR 745 at 750. 3. Simic (1980) 144 CLR 319 at 333.

In Schmahl [83] , in the passage approved in this Court, Sholl J had said:

I think it would not be right to lay down a rule that in every case where evidence of good character is given, the judge must give a direction as to the way in which it can be used. It is, of course, a different matter if the judge gives a wrong direction, as for example on the former somewhat more restricted basis stated in R v Bassett [84] , or if the prosecutor puts such a restricted view to the jury, when it would become the judge's duty to correct it. But, in general, non-direction with regard to good character evidence is not like non-direction with regard, say, to the evidence of an accomplice, where the jury without proper guidance might well misuse the evidence to the detriment of the accused. Ordinarily, if left without guidance, a jury would, I think, be inclined to use good character evidence in the way in which the High Court has said in Attwood v The Queen [85] , that it is to be used. But I would add that if counsel for an accused person asks for such a direction it would be wise for a trial judge to give it, lest it be afterwards suggested that in the circumstances of the particular case some less effective use of the evidence may have been made than the accused was entitled to expect. Personally, I always give the direction where evidence of good character is given. 1. [1965] VR 745 at 750. (In that case, the Full Court treated the value of evidence as to good character apparently as limited to consideration by the jury of whether it put a different complexion on the facts from that which they might bear without such evidence.) 2. [1952] VLR 535 at 541. 3. (1960) 102 CLR 353 at 359.

The question then is whether the Court should depart from the Australian position established in Simic [86] . There should be no such departure. The Court was referred to the treatment by the majority of the New Zealand Court of Appeal in R v Falealili [87] of the decision of the House of Lords in R v Aziz [88] with respect to the need for standard rules of practice.

1. (1980) 144 CLR 319. 2. [1996] 3 NZLR 664. 3. [1996] AC 41.

78. So-called standard rules of practice too readily ossify into rules of law which are difficult to vary or displace. In McDermott v The King [89] , Dixon J described as "a process that is not unfamiliar" the "growth of rules of practice and their hardening so that they look like rules of law". The line of cases, including McDermott itself, respecting the discretion to exclude confessional statements, which eventually fell for consideration in R v Swaffield [90] , illustrates the point.

1. (1948) 76 CLR 501 at 513-514. 2. (1998) 192 CLR 159.

79. I prefer as consistent with the established position in this Court, the statement of Thomas J in his dissenting judgment in Falealili [91] :

[I]f the evidence of the accused's good character is both probative and relevant the Judge will, almost as a matter of course, direct the jury as to its significance in summing up the defence case. It would be unfair not to do so. If, on the other hand, the purported character evidence is lacking in probative force and of remote relevance to the charge in issue, the Judge may decide that a good character direction is not warranted. Or the Judge may consider that it would be prudent to proffer a good character direction, but then to qualify it in order to put it in perspective having regard to the circumstances of the case. To proscribe that, whenever character evidence is adduced or elicited, a good character direction should be given and that it must generally embrace both the credibility and propensity limbs of the direction is an unnecessary fetter on that discretion.

In some cases the good character of the accused may be an integral part of the defence. A number of reputable persons may have testified as to the accused's character. In other cases the so-called evidence of good character may be little more than a passing reference, included by defence counsel, perhaps, simply because there is no other defence. In other cases the established facts of the case may itself indicate that, irrespective of how unblemished the accused's reputation may be, he or she can barely be described as a person of good character. Because the circumstances will vary greatly it is not possible to lay down comprehensive guidelines as to when and how the Judge's discretion should be exercised. Nor is it desirable to do so. Unless guidelines are treated as being just that, guidelines and no more, they could themselves inhibit the exercise of a Judge's discretion to do what is most appropriate having regard to the facts of the particular case.

80. The evidence as to the character of the appellant did not require the giving of a direction further to that which had been given.

82. "[I]n recent years there has been a veritable sea-change in judicial thinking in regard to the proper way in which a judge should direct a jury on the good character of a defendant" [92] . This remark of Lord Steyn identifies the central issue in this appeal which comes from the Court of Criminal Appeal of the Northern Territory [93] . What direction does the common law of Australia require a judge to give to a jury in a criminal trial where evidence is adduced which shows that the accused was of good character before the alleged offence?

1. R v Aziz [1996] AC 41 at 50. 2. Unreported; Court of Criminal Appeal (NT); 20 June 1997, per Gallop J (Martin CJ and Angel J concurring).

83. The "dramatic change" [94] which has occurred in England has attracted judicial [95] and academic [96] critics. One judge has observed that, ever since the change began thirty years ago [97] there has been trouble [98] . An academic commentator on the directions which English law has taken has described them as "preposterous" [99] , a "curiosity" [100] , even resulting in directions which cause one to doubt one's own sanity [101] . However, the trend of authority in England was recently endorsed by the House of Lords [102] . It has gathered majority support in the New Zealand Court of Appeal [103] and has staunch academic defenders as well [104] . This appeal affords an opportunity to this Court to declare the law applicable in Australia.

1. R v Vye [1993] 1 WLR 471 at 474; [1993] 3 All ER 241 at 243, per Lord Taylor of Gosforth CJ. 2. R v Falealili [1996] 3 NZLR 664 at 668-676, per Thomas J. 3. Munday, "What Constitutes a Good Character?" [1997] Criminal Law Review 247, at p 251 . 4. Beginning with R v Bellis [1966] 1 WLR 234; [1966] 1 All ER 552. 5. R v Wood [1996] 1 Cr App R 207 at 218, per Staughton LJ. 6. Munday, "What Constitutes a Good Character?" [1997] Criminal Law Review 247, at p 258 . 7. Munday, "What Constitutes a Good Character?" [1997] Criminal Law Review 247, at p 256 . 8. Munday, "What Constitutes a Good Character?" [1997] Criminal Law Review 247, at 258. 9. R v Aziz [1996] AC 41. 10. R v Falealili [1996] 3 NZLR 664. 11. eg, Ligertwood, Australian Evidence, 2nd ed (1993), p 129, par [3.60].

84. Mr Roy Melbourne (the appellant) was found guilty of murder following his trial in the Supreme Court of the Northern Territory. There was no contest that his were the acts which caused the death of the deceased. The real issue in the trial was whether the appellant had established diminished responsibility [105] . If the appellant could establish diminished responsibility, the jury were entitled to return a verdict of manslaughter. Once convicted of murder, the appellant had to be sentenced to life imprisonment - as indeed he was [106] .

1. Criminal Code NT, s 37. 2. Criminal Code NT, s 164.

At the close of the trial, the appellant's counsel requested that the judge give the jury directions to the effect that the appellant was a person of "good character" as that expression is used in this context. The evidence received at the trial established that the appellant had no criminal record other than for a twenty-year-old breathalyser offence; that he was not "adversely known to the police"; and that he was described by those who knew him as "amiable", "quiet", "gentle", "well behaved" and never "aggressive". No evidence of "bad character" had been tendered by the Crown in rebuttal. The trial judge accepted, and directed the jury, that the appellant could be regarded as a person of "good character". However, although she was expressly asked to give directions of the significance of that fact for the propensity of the appellant to commit the murder charged and for his credibility in out-of-court statements, the direction eventually given was confined to propensity. It gave no instruction to the jury on how they could use the established evidence of good character in evaluating the appellant's credibility.

Counsel for the appellant at the trial each swore affidavits which were read without objection in the Court of Criminal Appeal. They offered excuses for the failure to seek further direction concerning the issue of credibility. Leading counsel stated that, when the judge agreed, over the opposition of the Crown, to give the direction on good character, he assumed that the direction would cover both aspects, as requested, and was distracted. Because no attempt was made to question counsel's excuses, the inference should be drawn that the failure to make further objection was not deliberate or tactical. The point was properly reserved at the trial. The questions are thus posed as to whether the direction sought ought to have been given on both aspects and, as it was not, whether this caused a substantial miscarriage of justice requiring a retrial. To answer these questions it is necessary to identify the applicable rule and then to consider whether, if it was not observed in this trial, the appellant's conviction of murder should nonetheless be confirmed by the application of the "proviso" [107] .

Before identifying the controversies raised by the appeal, it is useful to narrow the contest before this Court. This can be done by noting a measure of common ground between the appellant and the Crown.

The appellant accepted that his case was complicated by the criticisms made during the trial concerning some of the statements which he had made to the police. For example, the appellant had told the police that he had cut his finger when cutting up a chicken. It was clear from the evidence that the cut actually occurred when the appellant stabbed the deceased. Even the appellant's case, therefore, accepted that some of his statements out of court to the police were false. Yet the appellant submitted that, given the scepticism with which a jury might examine claims of diminished responsibility, estimates of his general credibility were essential to the jury's assessment of his statements relevant to his state of mind. A distinction was drawn between deliberately dishonest out-of-court statements (for the appellant gave no evidence at his trial) and statements which, although honestly made, demonstrated and reflected an amnesic, confused state of the appellant's mental processes at the time of the killing.

The Crown accepted that the trial judge had been correct to conclude that the appellant had established that, prior to the killing of the deceased, he had a "good character". The Crown was even prepared to concede that a direction concerning the appellant's credibility was "technically applicable". Its argument was that (a) the giving of such a direction was discretionary, not mandatory; (b) the failure of counsel to seek further direction when the partial instruction on good character was given was fatal; and (c) in any case, the statements of the appellant to which a credibility direction would be relevant were made out of court and, for the most part, were inadmissible hearsay used by medical experts to ground their opinions. This meant, so the Crown submitted, that no miscarriage of justice had occurred which required a retrial. Whatever might be the case where an expert expresses an opinion based on direct observations and grounded in that expert's particular discipline [108] , in this case the expert opinions for both sides rested on out-of-court statements of the appellant which were never proved by sworn testimony. Such statements were therefore not properly admissible [109] . The opinions based upon them were not strictly available. To the extent that the appellant had enjoyed the benefit of those expert opinions, it was an advantage to which he was not strictly entitled. As such evidence went to his defence, on an issue upon which he bore the onus of proof, the want of a direction on credibility occasioned no injustice. Once the appellant failed to give evidence confirming the facts on which the expert opinions were based, any use of those opinions was in the nature of a windfall for the appellant.

1. R v Schafferius [1977] Qd R 213. 2. cf Ramsay v Watson (1961) 108 CLR 642 at 649.

90. It is a measure of the attention which has been given to the subject of "good character" in the context of criminal trials that numerous questions have been raised in the books. Many of them were debated in this appeal.

• Does "good character" exist at all? Is it an outmoded or antiquated [110] notion of morality and human propensity which has been overtaken by psychological experimentation and understanding, and which should no longer be reflected in the directions which judges give to contemporary juries?

• Are the considerations of propensity to criminal conduct and creditworthiness in statements made, whether on oath or otherwise, separate notions or different facets of the one idea that some people are prone to commit crimes and tell lies in ways that are predictable from their past conduct whilst others are not [111] ?

• Is it necessary, or appropriate, to draw a distinction between particular categories of crimes, such that evidence of good character or the absence of convictions will be treated as relevant to, say, a crime of dishonesty, but not necessarily to an unpremeditated, spontaneous crime of a sexual or other nature? Or is good character a badge of good citizenship which the law acknowledges, for policy reasons, so that it stands the accused in good stead when faced by a criminal accusation which goes to trial?

• Is the absence of prior convictions, without more, a demonstration of "good character" entitling an accused to a direction on that ground? Or does the law require more substantial and affirmative evidence to demonstrate the large and elusive concept of a person's "character" from which inferences about that person's conduct or credibility may safely be drawn by a jury?

• Are "good character" and "reputation" [112] , which are sometimes used interchangeably in the cases, the same notion? Or does "good character" refer to inner qualities of the accused which may or may not be reflected in that person's public reputation?

• Where evidence of good character is given, is the judge conducting a criminal trial bound as a matter of law [113] , or as a rule of practice [114] , to instruct the jury on the use to be made of it? Does such an obligation arise only in cases where a request for such a direction is made [115] ? Is it an obligation from which the judge is relieved if the proof of discreditable conduct or of lies would make the giving of a direction absurd or confusing in the circumstances [116] ? Or is it a matter for judicial discretion, having regard to the "feel of the case" and the need for particular assistance to the jury in the special circumstances of the trial [117] ?

• Is the direction in the nature of a warning, based on the experience of the law, akin to other warnings required of judges [118] ? Is it an illicit judicial comment on the facts which are the province of the jury to whom matters of common sense concerning the relevance of "good character" can safely be committed [119] ? Or, as a matter of policy, do inconsistent rulings by trial judges [120] , the proliferation of appeals on this point [121] , and the practicalities of framing standard judicial directions for inclusion in Judicial Bench Books [122] require that directions about "good character" be given in every case so as to avoid a justifiable sense of grievance at their omission and unnecessary appeals and retrials as a consequence [123] ?

• Is the credibility aspect of the direction on "good character" now an established part of the law [124] ? If so, is it limited in its application to evidence given on oath? Does it extend to statements made in the formal situation of a police interview? Or in informal answers to police? In statements given to or used by experts whom the person knows, or should know, may be called in evidence at the trial? Or generally?

• Given the developments in the common law which have occurred in other countries since this Court last considered the "good character" direction [125] , does the law in Australia now require judges to observe the law or practice which is established by judicial authority in other jurisdictions? Or are the directions to be given more properly matters to be left to the discretion of trial judges so that a more general reform of the law (if any) is left to the legislature [126] ?

1. Munday, "What Constitutes a Good Character?" [1997] Criminal Law Review 247, at p 248 . 2. Mendez, "The Law of Evidence and the Search for a Stable Personality", Emory Law Journal, vol 45 (1996) 221. 3. Sometimes "good fame and character" are referred to in statutory language suggesting that "fame" or "reputation" are different from character, being the external appreciation of internal qualities. See, eg, Legal Profession Act 1987 NSW, s 9; Re B [1981] 2 NSWLR 372 at 380; Wentworth v NSW Bar Association (unreported; Court of Appeal (NSW); 14 February 1994) at 2. 4. R v Aziz [1996] AC 41; R v Vye [1993] 1 WLR 471; [1993] 3 All ER 241. 5. R v Falealili [1996] 3 NZLR 664. 6. See Sholl J's comment in R v Schmahl [1965] VR 745 at 750. See also R v Falealili [1996] 3 NZLR 664 at 668, per Thomas J. 7. R v Aziz [1996] AC 41 at 53. 8. See, eg, R v Falealili [1996] 3 NZLR 664 at 668, per Thomas J, citing Devlin, Trial by Jury (1966), p 120: "The trial is not simply a trial before a jury; it is a trial before a Judge and jury requiring a "compounding of the legal mind with the lay" "; cf R v Schmahl [1965] VR 745 at 750. 9. eg, Longman v The Queen (1989) 168 CLR 79; cf R v Falealili [1996] 3 NZLR 664 at 669, per Thomas J. 10. R v Schmahl [1965] VR 745 at 750; cf Simic v The Queen (1980) 144 CLR 319 at 333-334. There are other instances where judges enter into the jury's role in various ways concerning the facts. The explanation of the judicial function in such cases is controversial. See, eg, Glass, McHugh and Douglas, The Liability of Employers in damages for personal injury, 2nd ed (1979), pp 209-210. 11. R v Aziz [1996] AC 41 at 53. 12. R v Vye [1993] 1 WLR 471 at 474; [1993] 3 All ER 241 at 243. 13. A matter mentioned by Thomas J. See R v Falealili [1996] 3 NZLR 664 at 673. 14. R v Vye [1993] 1 WLR 471 at 475; [1993] 3 All ER 241 at 244. 15. It was not mentioned as an aspect of "good character" in Attwood v The Queen (1960) 102 CLR 353. 16. In Simic v The Queen (1980) 144 CLR 319 at 333. 17. cf R v Aberg [1948] 2 KB 173; R v Smith [1971] Crim LR 531.

91. Other problems which I have not mentioned pose difficulties for a trial judge asked in a criminal trial to give a direction to the jury on good character. One, not raised by this appeal, concerns the direction which is appropriate in a joint trial of several accused, some only of whom can establish good character [127] .

1. R v Vye [1993] 1 WLR 471 at 478-479; [1993] 3 All ER 241 at 247-248.

92. Sufficient has been said to indicate the nature of the controversies and the desirability of clear guidance to trial judges so as to dispel uncertainty and to avoid unnecessary appeals.

93. The issue presented is not determined by the existing authority of this Court. Certainly, the Court has from time to time considered the law governing good character evidence. In Attwood v The Queen [128] the availability of evidence of "the favourable character" which an accused bears was mentioned. However, this was done in the context of elucidation of a statutory provision [129] forbidding, in certain circumstances, questioning of an accused which would tend to show that he or she was "of bad character". Too much weight should not be attached to the remark in Attwood that an accused "may adduce evidence of the favourable character he bears as a fact or matter making it unlikely that he committed the crime charged" [130] . The Court observed that evidence of good character "is regarded as really bearing on the probability or improbability of guilt" [131] . However, the lack of mention of the relevance of such evidence to the credibility of the accused may be understood when regard is had to the issues in that case and to the developments in case law since Attwood was decided.

1. (1960) 102 CLR 353. 2. Crimes Act 1958 Vict, s 399. 3. Attwood (1960) 102 CLR 353 at 359. 4. Attwood (1960) 102 CLR 353 at 359, attributing the principle to R v Rowton (1865) Le & Ca 520 at 530 [ 169 ER 1497 at 1502].

94. Some hint of those developments may be found in the description by Barwick CJ in Donnini v The Queen [132] of the "settled policy of the law" requiring, unless sanctioned by recognised exceptions to that rule, the exclusion of evidence that an accused person has previous convictions or is of bad character. Barwick CJ instanced as one of the illicit purposes to which "bad character" evidence might be put "the exposure of that character where the accused's credit is involved" [133] . In such a case "bad character" evidence would be "susceptible of use by a jury as indicating a propensity for criminal behaviour" [134] . Because an accused would, whether in or out of court, commonly deny guilt of the charge, issues of credibility and propensity to commit the offence are often (as Barwick CJ recognised) intermingled.

1. (1972) 128 CLR 114 at 123. 2. Donnini (1972) 128 CLR 114 at 123. 3. Donnini (1972) 128 CLR 114 at 123.

95. In Simic v The Queen [135] , the Court rejected a submission that a trial judge, who had not been asked to do so, had erred in failing adequately to instruct the jury in a criminal trial on the use to be made of evidence about the good character of the accused. The judge had mentioned that the accused had no prior convictions but had said nothing else. After adverting to Attwood, the Court observed that the treatment of the subject of good character in that case "did not purport to be a full statement of the law" [136] . In Simic it concluded that it was unnecessary to discuss the matter further. Nevertheless, the proposition that a direction on good character had to be given in every case was rejected as being based on "no rule of law".

1. (1980) 144 CLR 319. 2. Simic (1980) 144 CLR 319 at 333.

96. The Court in Simic [137] expressed agreement with the following passage in the reasons of Sholl J in the Full Court of the Supreme Court of Victoria in R v Schmahl [138] : "[I]f counsel for an accused person asks for such a direction it would be wise for a trial judge to give it." Sholl J observed in Schmahl : "Personally, I always give the direction where evidence of good character is given." [139] No more than Attwood does Simic represent a full treatment by this Court of the rules governing judicial directions on the use to be made of evidence of the accused's good character. Since the "dramatic change" [140] which has occurred in other common law jurisdictions and following a greater appreciation of the many questions that arise, it is necessary to reconsider the opinion offered in Simic about what the law requires. The Crown did not suggest that Simic bound this Court to approach all cases on the footing that the giving of a direction on good character was, as a matter of law, totally within the discretion of the trial judge. This appeal was argued on the basis that developments which have occurred elsewhere, and appreciation of the issues raised, now required this Court to say whether the provision of a direction is a matter for judicial discretion or, as a matter of law or practice, is a judicial obligation.

1. (1980) 144 CLR 319 at 333. 2. [1965] VR 745 at 750. 3. R v Schmahl [1965] VR 745 at 750. 4. R v Vye [1993] 1 WLR 471 at 474; [1993] 3 All ER 241 at 243.

97. To understand the present law, it is useful to know its history. Originally, English juries were self-informing. Independent and original knowledge of the facts, including of the character and reputation of the accused, was attributed to the jury [141] . As late as Bushell's Case [142] , Vaughan CJ is recorded as saying that "[t]he jury may know the witnesses to be stigmatiz'd and infamous, which may be unknown to the parties, and consequently to the Court". Only later did the view develop that the juror with personal knowledge must declare it in open court [143] , resulting, invariably, in disqualification of that juror from service in the trial so affected.

1. Thayer, A Preliminary Treatise on Evidence at the Common Law (1898), p 137. 2. (1670) Vaugh 135 at 147 [ 124 ER 1006 at 1012]. 3. Thayer, A Preliminary Treatise on Evidence at the Common Law (1898), p 139; cf Bennet and the Hundred of Hartford (1650) Sty 233 [ 82 ER 671].

98. Because until late in the nineteenth century, in Australia as well as England, an accused person was not competent to give evidence in a criminal trial, procedures developed, at first in capital cases as an indulgence defensive of life [144] and then more generally, to permit the accused to call evidence of good character. Naturally, such evidence was explained as going directly to the issue of the guilt of the accused of the crimes charged rather than the credibility of the accused. The latter was not originally seen to be an issue, because of the absence of sworn testimony from the accused [145] . In R v Stannard [146] , Patteson J explained:

[T]he object of laying it before the jury is to induce them to believe, from the improbability that a person of good character should have conducted himself as alleged, that there is some mistake or misrepresentation in the evidence on the part of the prosecution, and it is strictly evidence in the case. To similar effect were the remarks of Cockburn CJ in R v Rowton [147] cited in Attwood [148] . 1. In favorem vitae. See United Kingdom, Criminal Law Revision Committee, Eleventh Report: Evidence (General) (1972) Cmnd 4991, par 134; cf Munday, "Directing Juries on the Defendant's Good Character", Journal of Criminal Law, vol 55 (1991) 521, at p 522 . 2. Munday, "Directing Juries on the Defendant's Good Character", Journal of Criminal Law, vol 55 (1991) 521, at pp 522, 527 . 3. (1837) 7 Car & P 673 at 674-675 [173 ER 295 at 296]. 4. (1865) Le & Ca 520 at 530 [ 169 ER 1497 at 1502]. 5. (1960) 102 CLR 353 at 359; cf Taylor, A Treatise on the Law of Evidence, 4th ed (1864).

99. The rule permitting evidence of good character to be given was seen as an exception to the general principle which focused the attention of the trial on the crime alleged rather than upon the kind of person the accused was. The common law, for good reason, has always been, and still is, vigilant about the injustices which may be caused by propensity reasoning [149] .

1. Pfennig v The Queen (1995) 182 CLR 461 at 512-513; BRS v The Queen (1997) 191 CLR 275 at 320-321, 326; cf Phipson, The Law of Evidence, 3rd ed (1902), p 144: "Where character is tendered in proof or disproof of some other issue, it is usually excluded as irrelevant, whether in its sense of reputation or disposition " (emphasis in original); Smith and Holdenson, "Comparative Evidence: Admission of Evidence of Relationship in Sexual Offence Prosecutions - Part I", Australian Law Journal, vol 73 (1999) 432, at pp 434-436 .

100. Once the accused became able to give evidence in a criminal trial, the relevance of proof of good character for the accused's testimony was presented in sharp relief. Indeed, some judges, by the middle of this century, forgetting the history of the matter, suggested that good character was relevant "primarily [to the issue of] credibility" [150] and of little significance for a trial in which the accused elected not to give sworn evidence. This restrictive view was corrected in England in 1979 [151] when the two aspects of good character were emphasised: both propensity and credibility.

1. R v Bellis [1966] 1 WLR 234 at 236; [1966] 1 All ER 552 at 552. 2. R v Bryant [1979] QB 108.

101. By 1989, in England, directions about the relevance of previous good character both to propensity and credibility had become "conventional" [152] . Even before these developments had occurred in England, like changes were accepted in the practice of criminal trials in Australia. In R v Trimboli [153] , King CJ expressed three general propositions which became highly influential for judicial directions on good character in this country and in New Zealand. The first was that it was desirable "in all cases in which there is evidence as to the accused's good character" that a direction be given as to the use to which that evidence should be put. The second was that no particular form of words was necessary. The jury should be told to bear such evidence in mind "as a factor affecting the likelihood of the accused committing the crime charged". The judge might add "if he thinks it appropriate in the particular case" that the jury should consider the accused's previous good character "in assessing the credibility of any explanations given by him and, when he has given evidence, his credibility as a witness". The third principle which King CJ mentioned was that trial judges might remind the jury that people do commit crimes for the first time, a consideration with particular force in certain types of crime, notwithstanding evidence of past good character.

1. R v Berrada (1989) 91 Cr App R 131 at 134, per Waterhouse J. 2. (1979) 21 SASR 577 at 578. See also R v Fuller (1994) 34 NSWLR 233.

102. In the twenty years since Trimboli , still further developments have occurred. It gradually became accepted in England that a direction relevant to the credibility of the accused is required where that person has not given evidence but has made out-of-court statements, particularly to police [154] . That extension of the obligation, accepted by trial judges, was endorsed by the House of Lords in R v Aziz [155] in relation to answers given to police questions. A similar extension was accepted for New Zealand [156] . The law in Canada appears to be substantially the same [157] . Evidence of good character is admitted there for the dual purpose of demonstrating that the accused is not the type of person who would have committed the offence charged and to permit the jury to assess the accused's credibility where that is relevant. In Canada, a failure to instruct the jury as to the dual purpose of evidence of good character ordinarily constitutes appealable error [158] . The position in the United States is less helpful, being influenced by the provisions of the Federal Rules of Evidence [159] .

1. cf R v Bryant [1979] QB 108; R v Boyson [1991] Crim LR 274; R v Briley [1991] Crim LR 444; R v Vye [1993] 1 WLR 471 at 477; [1993] 3 All ER 241 at 247. 2. [1996] AC 41 at 52. 3. R v Falealili [1996] 3 NZLR 664 at 667. 4. R v McMillan (1975) 23 CCC (2d) 160 at 165, affd [1977] 2 SCR 824; R v Profit (1992) 11 OR (3d) 98 revd [1993] 3 SCR 637; R v Lizzi (1996) 2 CR (5th) 95 at 99. 5. R v Dees (1978) 40 CCC (2d) 58; R v Logiacco (1984) 11 CCC (3d) 374; R v Elmosri (1985) 23 CCC (3d) 503. 6. Rule 404(a). See also American Jurisprudence, vol 29, 2d, Evidence (1994), pp 399-417, ss363-374.

103. This history of the development of the law on good character in criminal trials demonstrates a consistent expansion of the judicial obligation to give directions to a jury in a criminal trial that evidence of past good character on the part of the accused may be used by them both to assess whether the accused is the kind of person who would commit the crime as charged (propensity) and to evaluate statements made both in and out of court that are relevant to the accused's guilt (credibility).

104. There are some cases where statute affects, or has affected, the directions which the judge must give [160] . It was not suggested that any statute applicable in the Northern Territory to a territory crime such as murder governed the law applicable in this appeal [161] . Nor in this case was there the kind of misdirection which required appellate correction in R v Murphy [162] . The essential issue is whether, in the circumstances of the case, the trial judge was required to give the direction on character evidence relating both to propensity and credibility as asked. To resolve that issue, it is necessary to clarify the applicable rule.

1. See, eg, Crimes Act 1900 NSW, s 412 (since repealed) considered in R v Murphy (1985) 4 NSWLR 42. 2. cf Evidence Act 1995 Cth and NSW, s 165. 3. (1985) 4 NSWLR 42.

105. A threshold point arises. Does "good character" exist in the form that has been assumed by centuries of the common law? Dictionaries suggest that "character" refers to the aggregate of qualities which distinguish one person from another, or to the "moral constitution" of a person. The etymology of the word, from a Greek word for an instrument used for engraving, suggests that "character" in relation to an individual refers to a permanent and unchanging pattern of the nature of the individual concerned. However, this reflects a now somewhat outdated view of complex psychological phenomena [163] . The belief that individuals are indelibly marked by an identifiable "character" has value in the law only so far as it is based on an assumption that such "character" has a predictive value, whether for good or bad. This notion is not only challenged by the fact that every first offender once had a "good character". It is also difficult to reconcile with modern psychological experimental literature. It appears to rest, like several common law rules of evidence, "on unstudied assumptions of human nature that generally have been rejected by those who have tested the actual effects of the rules of evidence on human behavior and decision-making" [164] .

1. cf McBride v Walton (unreported; Court of Appeal (NSW); 15 July 1994). At 21 of my opinion in that case I referred to Hilaire Belloc's mocking and sceptical poem The Statesman:

2. Okun, "Character and Credibility: A Proposal to Realign Federal Rules of Evidence 608 and 609", Villanova Law Review, vol 37 (1992) 533, at p 562 ; cf State Rail Authority (NSW) v Earthline Constructions Pty Ltd (In liq) (1999) 73 ALJR 306; 160 ALR 588.

106. To possess a predictive quality, as the law assumes in the case of good character, both in relation to propensity to crime and credibility, it must be hypothesised that the "character" of individual human beings demonstrates qualities which are sufficiently enduring and unvarying to be useful to a court. The law deals with myriad circumstances in which individuals are exposed both to similar and dissimilar situations [165] . But where a person does not have a stable personality or is exposed to new, special or extraordinary circumstances, the assumption that the person's conduct may be predicated on a previous absence of convictions, or even on a general reputation for, or existence of, good character, is doubtful [166] .

1. Mendez, "The Law of Evidence and the Search for a Stable Personality", Emory Law Journal, vol 45 (1996) 221, at pp 225-226 . 2. Reed, "The Character, Evidence Defense: Acquittal Based on Good Character", Cleveland State Law Review, vol 45 (1997) 345, at p 380 ; cf Smith and Holdenson, "Comparative Evidence: Admission of Evidence of Relationship in Sexual Offence Prosecutions - Part I", Australian Law Journal, vol 73 (1999) 432, at p 434 .

107. The previous assumption of lawyers, shared by some psychologists [167] , was that individual behaviour was comparatively stable under a variety of similar situations. Upon this view, behaviour arose from "certain attributes or mental structures called "traits" " unique to each individual [168] . That belief is now criticised as lacking empirical support [169] . Even "seemingly trivial situational differences" have been found to reduce correlations between accepted "traits" and the subject's conduct to zero [170] . Whilst human behaviour, whether conscious or unconscious, may not be "random or whimsical" [171] , and whilst characteristically predictable patterns may affect the behaviour of some individuals, even in quite different situations [172] , the intuitive assumption upon which judges of the common law have for centuries built notions of individual "character", predictive of criminal propensity and credibility, looks somewhat shaky when measured against modern psychological understandings and research. If this is so, the rules as to good character cannot be justified, still less extended, on the footing of available empirical evidence. There are, of course, repeat offenders as well as first offenders. But directions about evidence of "good character" in contemporary trials, if they are to continue, must rest on a basis of legal history and authority or upon an established legal policy, rather than demonstrable science.

1. Allport, Personality - A Psychological Interpretation (1937); cf Reed, "The Character Evidence Defense: Acquittal Based on Good Character", Cleveland State Law Review, vol 45 (1997) 345, at pp 353-355 . 2. Mendez, "The Law of Evidence and the Search for a Stable Personality", Emory Law Journal, vol 45 (1996) 221, at p 227 . 3. Mischel, Personality and Assessment (1968), p 147. 4. Mendez, "The Law of Evidence and the Search for a Stable Personality", Emory Law Journal, vol 45 (1996) 221, at p 228 , citing Mischel, Personality and Assessment (1968), p 177. 5. Reed, "The Character Evidence Defense: Acquittal Based on Good Character", Cleveland State Law Review, vol 45 (1997) 345, at p 357 . 6. Mischel and Shoda, "A Cognitive-Affective System Theory of Personality: Reconceptualizing Situations, Dispositions, Dynamics, and Invariance in Personality Structure", Psychological Review, vol 102 (1995) 246.

108. The foregoing is reason enough to reject the notion, found in many English decisions, that "good character" is synonymous with the absence of prior criminal convictions. The latter may be an indication of the former; but it is not necessarily so. I agree with Henry J's remark in R v Falealili [173] that "there are logical difficulties with the proposition that an absence of previous convictions is in itself evidence establishing a person's good character. It may be a factor in assessing good character, but standing on its own it is generally neutral." Proof of an absence of previous convictions, without more, would not, therefore, attract a judicial obligation to give directions about "good character". Fairness and balance in a charge to a jury might warrant mention by the judge of that fact. But for a "good character" direction, more evidence would be needed. To the extent that this distinguishes the rule applicable in New Zealand from that applicable in England [174] , I prefer the former.

1. [1996] 3 NZLR 664 at 667, per Henry J, with whom Eichelbaum CJ, Richardson P and Neazor J concurred. 2. Munday, "What Constitutes a Good Character?" [1997] Criminal Law Review 247, at p 255 ; cf Munday, "Directing Juries on the Defendant's Good Character", Journal of Criminal Law, vol 55 (1991) 521, at p 533 .

109. Too much water has flowed under the bridges of the common law to permit the overthrow by this Court of judicial directions favourable to an accused who is shown, by evidence in a criminal trial, to be of good general character. Despite the dubious psychological foundations of the law's previous rules, abolition is for the legislature, not the courts. The Crown did not submit that the past law should be abandoned. It simply argued that its expansion should be checked. Yet despite telling criticisms of psychologists and legal scholars [175] , it is too late simply to return the issue of good character to juries, unaided and from their common sense, to make what they will of such evidence as is given and as establishes that, in the past, an accused has had no convictions and may otherwise be accepted as a person of good character [176] . By the same token the foregoing considerations sound a note of caution about the unnecessary expansion of instructions about good character. Certainly, judges should not be compelled to give juries meaningless or absurd directions [177] nor directions which would "bemuse them" [178] or insult their good sense.

1. Munday describes the position reached in England as "risible". See Munday, "Directing Juries on the Defendant's Good Character", Journal of Criminal Law, vol 55 (1991) 521, at p 533 . 2. cf R v Falealili [1996] 3 NZLR 664 at 668, per Thomas J. 3. R v Aziz [1996] AC 41 at 53. 4. R v Falealili [1996] 3 NZLR 664 at 668, per Thomas J.

110. In response to the Crown's submission that the Court should leave it to the discretion of trial judges to decide whether instruction was or was not needed on the use of evidence of good character and on how to fashion the instruction, if given, in a way apt to the facts of the case, the appellant urged that this Court should follow the authorities in England [179] and New Zealand [180] and require a direction in every case in which evidence of good character is left before the jury, and then on each aspect of the matter (propensity and credibility) that may be relevant.

1. R v Aziz [1996] AC 41; R v Vye [1993] 1 WLR 471; [1993] 3 All ER 241. 2. R v Falealili [1996] 3 NZLR 664.

111. I acknowledge the strong reasons which the Crown advanced as to why this Court should hold back from expressing a general rule. The infinite variety of fact situations, illustrated by the peculiarities of credibility in the present case, speak for a degree of caution. So does the criticism just voiced concerning the intellectual foundations for the notion of "good character" and the objective reliability of its predictive value for criminal behaviour. So also does the oft-expressed trust and faith which the administration of criminal justice places in the jury and in their capacity to use their common sense in evaluating evidence. It has been suggested that the imposition of a rule obliging directions to be given in every case indicates not only distrust of juries; it also shows, so it was said, a "certain distrust of judges" [181] fairly and in a balanced way to call to the notice of juries matters of fact which stand to the credit of the accused, including the accused's past good character.

1. cf Munday, "What Constitutes a Good Character?" [1997] Criminal Law Review 247, at p 258 .

112. Whilst allowing that these arguments have weight and carry the majority in this Court, a number of considerations have persuaded me, on balance, to the opinion that the "sea-change" which has come over English law in this regard [182] and has now substantially been accepted as the law in New Zealand [183] and probably Canada [184] should also be accepted in the expression of the law of Australia. Let me explain why.

1. R v Aziz [1996] AC 41 at 50. 2. R v Falealili [1996] 3 NZLR 664. 3. The reason was explained by McLachlin J in R v Seaboyer (1991) 83 DLR (4th) 193 at 263 as being related to the extreme caution of the common law "founded in the fundamental tenet of our judicial system that an innocent person must not be convicted".

113. First, the trend of judicial authority in Australia has undoubtedly been towards upholding an accused's entitlement to have a direction where there is evidence of good character so that judicial authority and assistance are added to the pleas of counsel to the jury in that regard [185] . What was always regarded as good practice by leading judges has increasingly settled into best practice and is now reflected in Judicial Bench Books [186] . In this situation, omission of the direction about good character may result in a justifiable sense of grievance on the part of an accused who has demonstrated good character as that expression has come to be accepted by the courts.

1. See, eg, R v Trimboli (1979) 21 SASR 577; R v Gillard (unreported; Court of Criminal Appeal (NSW); 15 July 1991). 2. cf R v Falealili [1996] 3 NZLR 664 at 673.

114. Secondly, because (statute apart) the trial judge may be bound to remind the jury of the principal arguments of the prosecution, the law also requires that he or she bring to their notice, in a charge which is fair and balanced, matters of significance which favour the accused. According to accepted principles, evidence of good character is one such matter [187] . Circumstances may exist where the provision of the direction would not be warranted in the very peculiar facts of the particular trial. Where to give such a direction would be absurd or an affront to common sense, the common law would not impose that requirement on a judge. But ordinarily a charge to the jury which omitted reference to established evidence of good character would be unbalanced. The failure of the judge to explain how the evidence might be used by the jury would result in an incomplete direction.

1. R v Aziz [1996] AC 41 at 51; cf R v Falealili [1996] 3 NZLR 664 at 667.

115. Thirdly, although it cannot be said that Australian courts have been "inundated" [188] by appeals concerning judicial directions on good character, a number of such appeals have been heard. It is desirable that the obligations of best judicial practice should not depend unnecessarily on the inclinations of the particular judge who presides at the trial any more than on the skill and experience of the accused's advocate, which every court knows can be variable. Although there are disadvantages and risks in establishing a "clear-cut rule", for the avoidance of accidental injustice, unnecessary appeals, costly retrials and uncertainty, the recognition of a general rule represents the best and clearest policy [189] . It avoids any suggestion that the availability of the direction depends on a judicial "lottery" [190] . It leaves the trial judge in no doubt as to his or her duty. I see no advantage in describing this obligation as one of practice. Whilst the law of good character remains as it is, the judicial duty to give directions on the subject is an aspect of that law. Too much rigidity in judicial obligations in criminal and other jury trials is a burden, it is true. But the other side of the coin is judicial idiosyncrasy, variance and individual inclination. Too much of the latter will diminish the reality of the rule of law and substitute judicial rule and sometimes judicial whim or prejudice. These dangers can be avoided, as the courts in England, New Zealand and Canada have recognised, by the adoption of a simple and obligatory judicial requirement which, once observed, banishes the leeway for complaint.

1. R v Vye [1993] 1 WLR 471 at 474; [1993] 3 All ER 241 at 243. 2. R v Aziz [1996] AC 41 at 52. 3. R v Vye [1993] 1 WLR 471 at 474; [1993] 3 All ER 241 at 243. See also Munday, "Directing Juries on the Defendant's Good Character", Journal of Criminal Law, vol 55 (1991) 521, at p 532 .

116. Fourthly, in partial answer to the psychological and other critics of the common law's approach to "good character", it is probably worth observing that, if centuries of judges of our legal tradition have made assumptions about the consistent predictability of human conduct, jurors have probably also reasoned, and still do, in the same way [191] . It was on that footing, rather than on the bases of principle, logic and consistency, that Wigmore thought the better course was to admit evidence of good character [192] . Once such evidence is before the jury, it is reasonable to expect the judge to assist the jury in the use to be made of it [193] .

1. Smith and Holdenson, "Comparative Evidence: Admission of Evidence of Relationship in Sexual Offence Prosecutions - Part I", Australian Law Journal, vol 73 (1999) 432, at p 434 citing research by the Australian Law Reform Commission. 2. Munday, "Directing Juries on the Defendant's Good Character", Journal of Criminal Law, vol 55 (1991) 521, at p 532 . 3. R v Bruce [1975] 1 WLR 1252 at 1256-1257; [1975] 3 All ER 277 at 279.

117. Fifthly, in so far as, on occasion, the requirement of a judicial direction may seem over-generous to the accused, two answers may be given, one of principle and the other of practicality. To the extent that character evidence and an obligatory direction may appear "lopsided", that must be seen as yet another instance of the way the common law strives to avoid wrongful convictions by offering numerous protections favourable to the accused [194] . To the extent that it requires a direction which may seem unrealistic in the circumstances of the particular case, jurors can be expected to view it with "some scepticism" as "at worst [a manifestation of the] law's proclaimed solicitude for those accused of crime" [195] .

1. Ligertwood, Australian Evidence, 2nd ed (1993), p 129, par [3.60]. 2. Munday, "Directing Juries on the Defendant's Good Character", Journal of Criminal Law, vol 55 (1991) 521, at p 536 .

118. The trial judge in this case was therefore obliged to give a direction on good character. The obligation arose not because the appellant had no relevant prior convictions but because additional affirmative evidence was received at the trial which suggested that, before the killing, the appellant was what the law would classify as a person of "good character". The obligation was one incumbent upon the judge. But as it happened, in this case, it was reinforced by the application of counsel for the appellant.

119. Correctly, the judge at the trial gave part of the applicable direction, drawing to the attention of the jury the consideration that the past good character of the appellant might make it less likely in their view that he would act in the criminal way inherent in the charge of murder. However, although asked, the judge omitted to direct the jury that they might also take the established "good character" into account in assessing the credibility of the appellant's out-of-court statements to police and medical experts.

120. The development of this body of law requires some modification of the propositions which King CJ stated in Trimboli , although, in my view, as modified, these continue to give wise guidance to trial judges in Australia. The propositions which, in my opinion, should now be stated are:

1. In all cases in which there is evidence as to the accused's good character, a direction must be given by the judge as to the use to which that evidence may be put by the jury. Unless in the particular circumstances of the case doing so is unnecessary, or would be unwarranted for reasons which the judge gives, the directions on good character must relate both to (a) the way in which that evidence may be considered by the jury to make it less likely that the accused committed the offence charged (propensity) and (b) the reliance which the jury may place upon any evidence which the accused may have given in the trial and any other statements made by the accused out of court whether to police or others which come to the attention of the jury during the trial (credibility).

2. No particular form of words is necessary. However, the directions should convey to the jury that they should bear in mind the accused's previous good character when considering whether they are prepared to draw from the evidence the conclusion of the accused's guilt or a conclusion that the accused's evidence or relevant out-of-court statements are false. The jury are entitled to conclude that a person of established good character may be less likely to commit the crime charged or to make false statements relevant to guilt of that crime.

3. Because of the variety of the circumstances in which the need will arise, the directions must be tailored to meet the particular circumstances of the case. In an appropriate case, the judge will be at liberty to remind the jury that people do commit crimes for the first time and that evidence of previous good character is not a defence in itself and cannot prevail against evidence of guilt which, notwithstanding the accused's previous good character, the jury find to be proved. The judge may comment on the good character evidence and any rebutting evidence, in a fair and balanced way, including in relation to its significance or lack of significance in the circumstances of the particular case. Generally, however, such directions and comments should be brief because it can safely be left to the jury to apply their common sense to such matters.

121. I would leave to a future case the directions proper to a trial of co-accused, some only of whom are entitled to the benefit of a judicial direction concerning "good character". Until that question is presented for determination, it would be wise for Australian courts to follow the course adopted in England [196] and New Zealand [197] .

1. R v Vye [1993] 1 WLR 471 at 478-479; [1993] 3 All ER 241 at 247-248; R v Aziz [1996] AC 41 at 54. 2. R v Falealili [1996] 3 NZLR 664 at 667.

122. By the test of the foregoing rules, the directions given to the appellant's jury were inadequate. They omitted an important element in the explanation of the use which the jury might make of the evidence of the appellant's good character. No barrier of a procedural kind should prevent the appellant's relying on this point, which was raised at the trial. There was a misdirection. But does it require a retrial?

123. The Crown argued that the failure to give a direction on the "credibility" aspect of good character may not necessarily constitute a miscarriage of justice in every case and did not do so in this case. The appellant submitted that whilst his credibility as a witness was not relevant (because he gave no evidence) his out-of-court statements to the investigating police and to the expert witnesses did give rise to issues about his credibility which required that a direction be given to the jury, as his counsel specifically asked.

124. Much of the Crown's response to this branch of the appellant's argument, which concerned the appellant's statements to, or as used by, the expert witnesses, addressed the legal inadmissibility of the evidence of such experts without proper proof or formal admission of the facts upon which their opinions were based [198] . The Court of Criminal Appeal attached great significance to this consideration [10] . It appears to have led to that Court's conclusion that "evidence of good character [may not] be used to support the credibility of the accused in interrogation by police or the truthfulness of the history given to experts as a foundation of their respective opinions".

1. Referring to Ramsay v Watson (1961) 108 CLR 642; R v Schafferius [1977] Qd R 213 at 217, per Wanstall A-CJ. 2. Referring to Paric v John Holland (Constructions) Pty Ltd (1985) 59 ALJR 844 at 846; 62 ALR 85 at 87; R v Turner [1975] QB 834 at 840; Trade Practices Commission v Arnotts Ltd [No 5] (1990) 21 FCR 324.

125. The position of the appellant's statements to the police is distinct and separate, for such statements may be admitted in evidence as truth of the assertions contained within them. However, so far as the appellant's statements to, or as used by, the experts were concerned, no objection was taken at the trial that the statements were inadmissible, and the opinions therefore without foundation. On the contrary, at the trial both the Crown and the defence proceeded on the footing that the opinions were to be considered by the jury.

126. Whether that amounted to the admission of the statements by both sides for the purpose of providing the foundation for the expert opinions or whether it constituted a waiver by the Crown of the objection now pressed, it is wholly unrealistic to divorce from consideration the statements made by the appellant used by those witnesses. The jury, without special instruction that they might take any different course, would scarcely have done so [11] . Accordingly, the credibility of the appellant in the statements which he made both to or for the experts and to the police was potentially relevant to the jury's evaluation of his defence of diminished responsibility. If the jury considered that the appellant was a liar, cunningly attempting to manipulate both police and experts after he had killed the deceased in rage, so as to persuade the jury to reduce his offence from murder to manslaughter, a different verdict would result from that which would follow if the jury concluded (assisted by judicial directions on the appellant's past good character) that both to police and experts (and in his answers to the "Lezak" test) the appellant had been honest and credible, although sometimes factually mistaken.

1. cf R v Sharp [1988] 1 WLR 7 at 15; [1988] 1 All ER 65 at 71, per Lord Havers, approved R v Aziz [1996] AC 41 at 49.

127. The passages in the prosecutor's final address to the jury, set out by Callinan J in his reasons, demonstrate that, in several respects, the Crown challenged the truthfulness of the appellant's out-of-court statements both to the police and upon which the expert opinions in his favour were based. In these circumstances it is impossible to suggest that a direction on the use which the jury might make of the relevance of good character, as that consideration affected the credibility of the appellant, could have had no substantial effect on the jury's deliberations. It is therefore impossible to say that the direction could not have affected the outcome of the trial on the issue of diminished responsibility, which was the real issue that was litigated, or that the accused's conviction was inevitable.

128. Sometimes, looking at the conduct of a trial taken as a whole, the issue of good character may be viewed as "very much a footnote to the case" [12] . Sometimes a court may conclude, although with "hesitation, indeed with reluctance" [13] , that the "proviso" should not be applied. But where a direction on credibility of evidence or out-of-court statements should have been given and was not, and the challenged statements concern an important and contested issue, it would be rare that the proviso would be applied [14] . If the only out-of-court statements to which credibility might be relevant were clearly self-serving assertions of the accused or were made in a social or like setting, different considerations might apply. But where, as here, they are statements made to police conducting a formal interview or in circumstances considered appropriate to be placed before and relied upon by experts called in the trial, such statements take on (whether in law, or practicality, or both) a potential significance for the jury's deliberations.

1. R v Levy [1987] Crim LR 48, at p 49; cf Munday, "Directing Juries on the Defendant's Good Character", Journal of Criminal Law, vol 55 (1991) 521, at p 529 . 2. As was the case in the Court of Appeal in R v Aziz [1996] AC 41 at 48. 3. R v Vye [1993] 1 WLR 471 at 480; [1993] 3 All ER 241 at 249.

129. It was for that reason that the appellant was entitled to have the jury instructed on the approach which they should take to his credibility based upon the fact that, before the trial, he was a person whom the jury could consider to be of good character. This was not a case where it was inevitable that the jury would reject the appellant's claim of diminished responsibility. The expert evidence on the subject was divided [15] . There was some circumstantial evidence which supported the conclusion favourable to the appellant. It cannot be said affirmatively that the appellant lost no real chance of a verdict of not guilty of murder but guilty of manslaughter [16] .

1. Professor Tiller and Dr Milton disputed the claim of diminished responsibility; Dr Vine, Dr Walton and Mr Taylor supported it. 2. Mraz v The Queen (1955) 93 CLR 493 at 513-514; Wilde v The Queen (1988) 164 CLR 365 at 371-372; Berry v The Queen [1992] 2 AC 364 at 381.

The conviction of the appellant of murder, and the mandatory sentence of life imprisonment which followed, necessitate close attention to the conduct of his trial and the directions of law which the trial judge gave during the trial. By the high standards demanded by our law in such matters, this is not a case where, notwithstanding established error, the verdict can stand. There must be a retrial. Thereby the law in Australia should be rendered into the same simple and uniform state in which it now appears in England, New Zealand and, seemingly, Canada. Clarity and certainty should replace the chance of an advocate's lack of vigilance and the inclinations of an individual judge's disposition.

The appeal should be allowed. The orders of the Court of Criminal Appeal of the Northern Territory should be set aside. In place of those orders, it should be ordered that the appeal to that Court be upheld, the conviction of the appellant set aside and a new trial had.

The appellant was tried and convicted in the Supreme Court of the Northern Territory on a charge of murdering his next door neighbour. The appellant did not deny that he had stabbed the victim; he asserted that he did so in a state of diminished responsibility [17] . The appellant contends (and contended in the Court of Criminal Appeal) that his trial miscarried because the trial judge did not instruct the jury that it could take account of the appellant's good character in deciding whether to accept that he had tried to tell the truth when he was interviewed first by police and later by experts who were called at his trial to give evidence of his psychiatric state. The appellant's belief in what he told police and the experts was said to bear upon his defence of diminished responsibility.

At the request of counsel who then appeared for the appellant, the trial judge had instructed the jury that:

you are entitled to consider the improbability of [the appellant] committing the instant offence, having a history of good character of some 61 years, and that this is of considerable significance. No exception was taken to this direction and, after the trial judge had completed her charge, trial counsel sought no additional direction about the use the jury might make of the evidence about the character of the appellant.

On appeal, however, both in this Court and in the Court of Criminal Appeal, it was submitted that a more extensive direction should have been given - one that canvassed the use that a jury might make of the appellant's good character in deciding whether to believe what he had told police or expert witnesses. An affidavit of trial counsel was received in evidence by the Court of Criminal Appeal in which he said:

I have read page 357 of the Appeal Book. I misheard Her Honour's direction as to character and believed that the full direction sought had been given. I did not make a tactical decision to accept Her Honour's direction on character evidence as sufficient. Trial counsel was not cross-examined and it is not possible to say how or why it was that he "misheard" what was said by the trial judge and yet believed that "the full direction sought had been given".

Much of the argument in this Court (and in the Court of Criminal Appeal) assumed that the directions sought at trial by counsel for the appellant extended to a direction that the evidence of good character might be taken into account by the jury in deciding whether to believe the accused's out of court statements. I do not accept, however, that the request made by trial counsel went so far. In his submissions to the trial judge about what directions should be given, counsel described what he called the "aspect of [the appellant's] credibility" as "probably of lesser significance" and explained to the trial judge why that was so. He said:

it would be my submission that the aspect of the improbability of committing the instant offence, having a history of good character for sixty-odd years, is of considerable significance.

The aspect of his credibility is probably of lesser significance, having regard to the nature of the interview itself that has been severely criticised by me as showing a lack of credibility, but, in any event, it is the primary aspect of the evidence, in my submission, that should be brought to the jury's attention and one which, in our submission, as a matter of law, he is entitled to.

I thought the most convenient repository of the law in relation to this is a decision of R v Murphy [18]

The situation in Murphy's case, obviously, was that it's the aspect of credibility which was the most significant. In my submission, in this case it's the reverse; it's the aspect of probability or inherent probability of the commission of the offence. (Emphasis added.)

Taken as a whole, the submission made at trial was limited to a submission that the jury should be directed about the use of evidence of good character in deciding whether the appellant had committed murder. Although trial counsel swore that "[a]t the conclusion of the trial [he] sought a full direction on character evidence from Her Honour the trial judge as to both issues and credibility", I do not accept that this was the effect of what he said to the trial judge.

The fact that the appellant's counsel did not seek a direction about the use of the good character of the appellant in assessing his credibility presents a considerable difficulty for the appellant in now maintaining that such a direction should have been given. It is as well, however, to leave these difficulties on one side and to consider some aspects of the underlying principles.

In deciding whether a trial miscarried for want of a proper direction to the jury it is essential first to identify what were the issues in the case. Only once the issues are identified can an appellate court decide if the jury should have been given some instruction that they were not. Especially is that so when the direction that it is said should have been, but was not, given is a direction about the jury's reasoning on questions of fact.

In this case, trial counsel told the trial judge that the appellant's character was significant because it affected "the aspect of probability or inherent probability of the commission of the offence". But as he said to the trial judge, he had submitted to the jury that the appellant's interview with the police "show[ed] a lack of credibility". By this, it seems that counsel meant that what the appellant had told police was not accurate. In his address to the jury trial counsel had said that although the appellant "was fair dinkum about trying to answer the police questions", he was "just lost" and he gave more than one example to the jury of what he described as "false memories" of the appellant. He also suggested to the jury that the appellant had let pass opportunities presented in the interview to improve a defence of diminished responsibility.

In these circumstances, what was the issue that required the trial judge to tell the jury that the previous good character of the accused could be taken into account in deciding whether to believe what he had said out of court? Counsel then appearing for the appellant had sought to persuade the jury that they should not accept the accuracy of what the appellant had said to police. And trial counsel did not seek to argue, at trial, that the appellant's character could be used to support the jury's acceptance of what the appellant had said to the expert witnesses. Trial counsel for the appellant did seek to persuade the jury that his client had tried to tell the truth to police but he did not seek to buttress that argument by any but the most passing and inconsequential reference to the appellant's previous good character.

Having regard to the way trial counsel for the appellant dealt with these matters in his address to the jury, I do not accept that there was any issue at the trial about the use of the appellant's previous good character as an aid to the jury deciding whether to accept the accuracy of what he had said out of court or deciding whether he had tried to tell the truth. That would be reason enough to conclude that this trial did not miscarry for want of the direction for which the appellant contends. It is as well, however, to go on to consider some other, more general questions relating to directions about the previous good character of an accused. But before turning to consider some of those issues, it is necessary to restate some basic propositions.

The task of directing a jury in a criminal case is never easy. It would be made no easier (and would serve no purpose) if trial judges were bound to give more, and more complicated, directions than the particular case requires. But the obligation of a trial judge, onerous as it is, does not extend so far.

The task of the trial judge was stated in Alford v Magee [19] in the following terms:

[I]t may be recalled that the late Sir Leo Cussen insisted always most strongly that it was of little use to explain the law to the jury in general terms and then leave it to them to apply the law to the case before them. He held that the law should be given to the jury not merely with reference to the facts of the particular case but with an explanation of how it applied to the facts of the particular case. He held that the only law which it was necessary for them to know was so much as must guide them to a decision on the real issue or issues in the case, and that the judge was charged with, and bound to accept, the responsibility (1) of deciding what are the real issues in the particular case, and (2) of telling the jury, in the light of the law, what those issues are. If the case were a criminal case, and the charge were of larceny, and the only real issue were as to the asportavit, probably no judge would dream of instructing the jury on the general law of larceny. He would simply tell them that if the accused did a particular act, he was guilty of larceny, and that, if he did not do that particular act, he was not guilty of larceny. It may be that the issues in a civil case tend, generally speaking, to be more complex than in a criminal case. But the same principle is applicable, and looking at the matter from a practical point of view, the real issues will generally narrow themselves down to an area readily dealt with in accordance with Sir Leo Cussen's great guiding rule. [20] . These principles are especially important in criminal cases. The directions that a trial judge gives the jury in a criminal trial must instruct the jury on only so much of the law as they need to know for the purposes of deciding the particular case that has been tried before them. It is neither necessary nor desirable that a judge's charge go further. 1. (1952) 85 CLR 437 at 466, per Dixon, Williams, Webb, Fullagar and Kitto JJ. 2. See also Holland v The Queen (1993) 67 ALJR 946 at 951; 117 ALR 193 at 200, per Mason CJ, Brennan, Deane and Toohey JJ.

It is trite to observe that the jury, not the judge, are the sole judges of questions of fact. But that does not mean that a trial judge can leave all questions of fact to the jury without giving them any directions. The trial judge in a criminal trial must instruct the jury about some matters that affect how they set about finding the facts. Thus in some cases the judge must warn the jury of dangers of which they must beware when they are considering the facts. Directions about the dangers of identification evidence [21] or about accepting uncorroborated evidence in some circumstances [22] provide ready examples. But it is always necessary to bear steadily in mind that it is the jury that decide the facts - not the trial judge. Especially is this necessary when the question is (as it is in this appeal) whether a trial judge is bound to direct a jury on some matter that touches how the jury finds the facts in the case. The warnings about factual issues that I have mentioned are given to the jury not just because they relate to one or more of the issues in the case but because, if they are not given, the jury may omit consideration of important matters (of which they may be unaware) and wrongly conclude that guilt has been demonstrated beyond reasonable doubt.

1. Domican v The Queen (1992) 173 CLR 555. 2. eg, Longman v The Queen (1989) 168 CLR 79.

Evidence of the good character of the accused raises different questions. The direction that the appellant submits should have been given here was not a direction that would warn the jury to avoid a false chain of reasoning.

On appeal to this Court, counsel for the appellant contended that the jury might properly have considered the previous good character of the accused in deciding what use they might make of the answers he had given to police or to interviewing experts. As I have indicated earlier, the short answer to this contention is that no argument of this kind was advanced at trial. But assuming that is not right, and the argument was raised, what use could profitably have been made by the jury of the evidence that was led about the appellant's character?

If the appellant's answers to police were said to reveal that he was confabulating, this may, perhaps, have had some significance in deciding what was his mental state at the time of the interview which took place soon after the killing and, by inference, his mental state at the time of the killing. The experts who examined the appellant looked at his record of interview with police and gave evidence of interviews and tests they conducted. Whether the appellant had told the police and the experts what he believed to be the truth or, instead, had tried to establish some false basis for a later plea of diminished responsibility might have been said to affect whether the opinions of the experts were well founded. But debate at the trial focused upon the validity or accuracy of the opinions expressed by the experts rather than on their factual basis. Indeed, no point was taken at trial about whether any sufficient evidentiary basis was established for the opinions of the experts. No point was taken about whether some or all of the material relied on by the experts was hearsay or original evidence. The appellant's belief in what he said did not become a significant issue. And, as counsel for the appellant made clear to the trial judge, it was contended that what the appellant had said to the police (and the experts) was objectively untrue.

How did the evidence of his character relate to any of these matters?

The appellant had a prior conviction for a drink driving offence but it was said in evidence that he was not "adversely known to the police" and was "amiable", "quiet", "gentle", and never aggressive. Accepting all of these descriptions of his personality as accurate, and assuming (contrary to the fact) that he had never been convicted of any criminal offence, what does any of that say about whether he was confabulating in an interview with police? What would any of it say to a suggestion that he was deliberately feigning mental illness? The short answer is, nothing.

The law restricts the evidence that may be given at trial about the good reputation or character of an accused person [23] . The rule in R v Rowton (that evidence of character may be given only by giving evidence of reputation within a particular group or neighbourhood) has been described as "difficult to apply, widely ignored and logically unsound" [3] . The law also restricts the circumstances in which evidence of bad character may be given [4] and it may be that the criticisms of the rule in R v Rowton have less application to these rules. But the restrictions that are imposed in relation to character evidence are imposed because evidence of the character of an accused (good or bad) may distract attention from the central question in the trial: whether the prosecution has established proof of the offence charged. In some cases (perhaps many) debating what the accused has or has not done on other occasions will be of little profit.

1. R v Rowton (1865) Le & Ca 520 [ 169 ER 1497]. 2. Cross on Evidence, 5th Aust ed (1996), p 537 (footnotes omitted). 3. Evidence Act NT, s 9(7). Some States have similar provisions, eg, Evidence Act 1929 SA, s 18(1)VI; Evidence Act 1977 Q, s 15(2); Evidence Act 1906 WA, s 8(1); Evidence Act 1910 Tas, s 85(1). The regime under the Commonwealth and New South Wales Evidence Acts varies in some respects but these differences need not be noted now.

151. Nevertheless, the fact that an accused is a person of good character may loom large at trial. It may be a very persuasive argument in the hands of the accused's advocate and may be very influential in the jury's deliberations. In some cases, it may lead the jury to conclude that they are not satisfied of the guilt of the accused. In at least some cases that may owe more to an appeal to emotion or prejudice than to any identifiable and logical process of reasoning.

152. The argument that an accused is of previous good character seeks to attribute a single qualitative description ("good") to an indivisible character. But people are not divisible into two classes: those who are good and those who are not. And the use that a jury may make of such evidence as is given about the previous character of an accused will vary greatly according to the circumstances of the case. It will vary according to what is said about the previous character of the accused and what relationship (if any) that has to the case that it is sought to make against the accused. The submissions of trial counsel acknowledged that in this case the evidence of the appellant's good character had little to say about his credibility; in my view it provided no assistance to the jury on any question of the appellant's credibility.

153. The appellant's submissions in this Court proceeded from the premise that a trial judge must, if asked, give the jury a direction that evidence of the previous good character of the accused may be used in two ways: first, as an aid in assessing the probability that the accused committed the offence and secondly, as an aid in deciding whether to believe what the accused has said in the witness box or, as in this case, out of court.

154. In Simic v The Queen [5] it was held that "[t]here is no rule of law that in every case in which evidence of good character is given the judge must give a direction as to the manner in which it can be used". Since Simic was decided, the Court of Appeal of New Zealand and the House of Lords have both considered what directions should be given about the way in which evidence of good character can be used. In R v Falealili [6] the New Zealand Court of Appeal held that, where evidence of good character is given, it should be the general practice that an appropriate direction as to good character is given. (The Court of Appeal held that the omission of such directions in the particular case had not led to any miscarriage of justice.) In R v Aziz [7] the House of Lords held that, subject to the exercise of a residual discretion (reserved, it seems, primarily for cases where to give the direction would offend common sense [8] ), a trial judge should prima facie give such directions in any case in which the accused is shown to be of previously good character.

1. (1980) 144 CLR 319 at 333, per Gibbs, Stephen, Mason, Murphy and Wilson JJ. 2. [1996] 3 NZLR 664. 3. [1996] AC 41. 4. R v Aziz [1996] AC 41 at 53, per Lord Steyn.

The adoption of the rules I have described in New Zealand and in England and Wales seems to have been influenced more by the desire for establishing a certain rule that is easy to apply than by considering what is the place of such directions in a trial. Certainty and ease of application are powerful arguments. If the rule is that a judge must (or nearly always must) give these directions the trial judge need make no greater decision than a decision about where, in the charge, the direction should fit. Adopting a general and all embracing rule of the kind spoken of in Falealili and Aziz will make it easy to see whether a trial has miscarried on this account. But certainty and ease of application must be considered against what it is that the direction achieves. There is no point in insisting that a trial judge must give such a direction in every case (or nearly every case) in which good character is established unless to do so assists in achieving a fair trial. And that directs attention to what is in issue at trial.

When is there an issue about the use of character evidence that will call for judicial direction of the jury? The simplest example is, of course, if prosecution and accused make contrary submissions to the jury about whether evidence of prior good character can be used by the jury in assessing the probability of the accused committing the offence charged or in assessing whether the accused should be accepted as having sought to tell the truth in statements he or she has made in or out of court. Clearly, in such a case the judge must tell the jury what is the true position in law: that the previous good character may be used in either or both of these ways. And even if there is no conflict between the parties in their submissions to the jury about how the evidence may be used, there may be occasions where it may be wise for the trial judge to draw the matter to the attention of the jury. So, to take a common example, if an accused of previously undoubted honesty in money matters is tried for an offence of fraudulently obtaining financial advantage, the judge may think it appropriate to draw the attention of the jury to the fact that prior good character may be thought, by them, to make it less likely that the accused acted with dishonest intent. But even in such a case, if no more is known than the bare facts of the case as I have described them, there is no requirement for the judge to give such a direction. Or, to put the matter another way, the absence of such a direction does not lead to the conclusion that the trial miscarried.

There is no reason to depart from the conclusion stated in Simic, namely, that there is no rule of law that in every case in which evidence of good character is given the judge must give a direction as to the manner in which it can be used. Of course, if a direction is given, it must be accurate. Ordinarily, however, unless the evidence that is led about the character of the accused has an immediate and obvious connection with an issue in the case, it is better that the judge say nothing of how the jury may use such evidence in reasoning to its conclusions beyond any restatement of counsel's arguments that may be thought necessary or desirable.

The evidence of the appellant's character did not, in this case, require the trial judge to give a further direction than she did. The appeal should be dismissed.

This is an appeal from the Court of Criminal Appeal of the Northern Territory. The appellant was convicted of murder by the Supreme Court of the Northern Territory at Darwin on 13 June 1996. The mandatory sentence of life imprisonment was imposed. The appeal raises a question whether, in what circumstances, and in what terms a character direction must be given in a criminal trial.

The appellant, who was sixty-one years old at the time of trial, had lived next door to the deceased for a number of years. There was little contact between them. Evidence was given by the deceased's granddaughter that, on 27 July 1995, the appellant came on to the deceased's verandah, swore at her and said "You were banging on the walls". The deceased was heard to scream. She was stabbed three times. Subsequently, a neighbour (Jack Gooch) came to the victim's residence, saw the appellant kneeling over the deceased and told him to leave her alone. The appellant said "Yeah, Jacko, all right". He was "very quiet" and walked away. He lay down on the grass and said "Call the police, Jacko". Subsequently he told Nancy Barnes, in a flat tone of delivery, "I did it because she just wouldn't stop". Dorothy Hinde heard him say in "sort of a mumble, a monotone" that "it serves her right". He lay on the grass "languidly" waving his arms in the air. He lay there until the police arrived and arrested him.

The police testified that when they saw him the appellant appeared to have been drinking and had a very glazed and distant expression, a "blank expression". He did not respond to questions and was unsteady on his feet. Detective Sergeant Rowbottam formed the impression that there was "something wrong" with him. Subsequently he was breathalysed (with a blood alcohol content of 0.136 per cent). He was spoken to by Dr Nitschke a few hours later and stated that he did not know why he had been arrested. He was questioned by police in a recorded interview the following day and repeated a claim of what was, in substance, amnesia.

There was evidence that for a considerable period before the killing, the appellant had been hearing loud banging noises in his unit. He stated to several people that he believed the noises were coming from the deceased's unit next door and believed that she was making the noises deliberately. There was evidence that, in fact, the noises were real noises caused by plumbing problems although for a period the doctors called for the defence had proceeded upon the basis that the appellant's complaints about the noises were entirely delusionary. The appellant was "despondent" and found the noises so unsettling that he had decided to move to Queensland. On the day before the assault he bought a bus ticket to Mt Isa for the next day and removed almost all of his furniture. The Crown did not contend that these arrangements were evidence of premeditation.

There was also evidence that the appellant had a history of depression, alcoholism and "binge" drinking, benzodiazepine abuse and poor health.

The defence case was that the appellant was at the relevant time acting in a state of diminished responsibility. Under s 37 of the Criminal Code NT, a person who has unlawfully killed another in circumstances which would have constituted murder but was

in such a state of abnormality of mind as substantially to impair his capacity to understand what he was doing or his capacity to control his actions or his capacity to know that he ought not do the act is guilty of manslaughter only.

The appellant bore the burden of proof in respect of the defence of diminished responsibility. He called three expert witnesses (Mr Taylor, Dr Vine and Dr Walton) who gave evidence that in their opinion the appellant was in a state of diminished responsibility at the time of the killing of the deceased. The Crown called two expert witnesses (Professor Tiller and Dr Milton) who were of a contrary opinion.

In summary, the defence expert witnesses gave evidence that the appellant suffered: (a) frontal lobe damage (caused by alcohol and benzodiazepine abuse and causing cognitive deficits); (b) clinical depression; and (c) a delusional disorder (that he was being persecuted by the deceased).

The expert witnesses for the defence gave evidence that they substantially based their opinions on a number of out-of-court assertions made by the appellant, including: (a) assertions by the appellant that he had no memory of the stabbing (there was some support for this from memory tests conducted by Mr Taylor); (b) assertions by the appellant that he believed that the deceased was banging on the walls of her residence in order to upset him and make him leave his flat: these were relied on by Dr Vine to support an opinion of delusional disorder and also by Dr Walton; (c) assertions by the appellant regarding his alcoholism: relied on by Mr Taylor and given considerable importance by him; relied on also by Dr Vine and Dr Walton; (d) assertions by the appellant that he had been using large quantities of benzodiazepines: relied on by Mr Taylor and Dr Vine; (e) assertions by the appellant of a history of insomnia, poor appetite, social withdrawal and despondency: relied on by Dr Walton to support a diagnosis of depression; (f) answers given to a "Lezak" test (designed to support the genuineness of the appellant's responses to other neuropsychological clinical tests conducted on him by Mr Taylor): based on these tests Mr Taylor expressed the opinion that the appellant suffered from a degree of organic brain damage affecting the frontal lobes of the brain; Dr Vine relied on Mr Taylor's tests as did Dr Walton.

The Crown medical witnesses were not able to examine the appellant before trial, and they were not provided with copies of any written opinions by the experts called for the defence. As a result of a ruling by the trial judge the Crown called the experts for the prosecution as part of the Crown case. This ruling was made in rejection of an application to the trial judge that, in the circumstances, the Crown should be permitted to call its evidence in reply.

There was no objective physical evidence to support a finding of frontal lobe damage. Two "CT" brain scans, which were regarded by the Crown witnesses as potentially significant supporting procedures had been carried out with negative results. The second of the scan results was obtained by one of the defence psychiatrists but it only came to the notice of the Crown during cross-examination late in the trial. The finding of frontal lobe damage depended entirely therefore upon the appellant's history as recounted by him to the experts, inferences to be drawn from his conduct at the relevant time as described by other witnesses, and testing carried out by the psychologist Mr Taylor.

Issue was taken with respect to the existence of the two major findings of frontal lobe damage and delusional disorder. As to the other disease or illness which was canvassed, clinical depression, there was a division of opinion whether it was present, and if it was, as to its severity.

The Crown case was that the accused had been harbouring a pent-up anger for some time about the victim's supposed noisy habits and persecution of him.

Her Honour the trial judge in the absence of the jury and before she had finished her summing up asked counsel whether they wished to raise any matters. This exchange then occurred:

Mr Van de Wiel: I do, your Honour, and that is this. The direction in terms of character evidence. It's my submission that in the course of this trial the issue of Mr Melbourne's character was canvassed throughout the trial with just about every witness we had presented, who was a lay witness, if I can use that term, and also with Mr Newman. On that basis it would be my submission that the aspect of the improbability of committing the instant offence, having a history of good character for sixty-odd years, is of considerable significance.

The aspect of his credibility is probably of lesser significance, having regard to the nature of the interview itself that has been severely criticised by me as showing a lack of credibility, but, in any event, it is the primary aspect of the evidence, in my submission, that should be brought to the jury's attention and one which, in our submission, as a matter of law, he is entitled to.

I thought the most convenient repository of the law in relation to this is a decision of R v Murphy [9]

The situation in Murphy's case, obviously, was that it's the aspect of credibility which was the most significant. In my submission, in this case it's the reverse; it's the aspect of probability or inherent probability of the commission of the offence.

I say that despite the fact that while I have submitted to the jury that it's quite appropriate for them to convict him of murder because, in my submission, what the jury are involved in trying here is culpability for murder and that that, as a matter of law, Mr Melbourne can rely on his character as showing that it is less probable that he would have committed that crime.

The aspect of the conviction of 1975 of exceeding .08, in my submission, does not really affect the issue of character because it was a statutory offence, it's character neutral in that sense and particularly in the history of Mr Melbourne.

My submission is that character is not an issue in the way that my learned [friend] suggested in this case. Control is the issue in this case. My learned friend ran the case on this basis, that Mr Melbourne did behave himself in the past, he was then controlled - and this is the medical evidence he relies upon - and he's not - and was not then diminished. He now is diminished and that's the issue the jury's to be concerned with, not whether or not he was of good character at other times, because the point is that, as my learned [friend] has just said himself, it's admitted that he did it; the only issue is what his state of mind was when he did do it. If he's out of control it doesn't matter what his character was ten years ago or twenty years ago or thirty years ago. So, to the primary basis on which the issue goes, as my learned friend would argue it, it's not applicable.

If it's said to be applicable to credibility, that will only be of relevance, your Honour, if he gave evidence. If my learned friend wants to have credibility an issue in what he says in his record of interview, in my submission, it's not relevant to credibility unless there is evidence given. I hear my learned friend murmuring about that, but that's as I understand it.

When it came to the point her Honour gave a direction in these terms:

The first matter is a direction I propose to give you in respect of what use you make of the evidence you heard during the course of these proceedings, of the good character of the accused, Roy Melbourne.

I do not propose to go back and remind you again of that evidence, but there has been evidence of course from witnesses relating to his not being a violent or aggressive person, and references to the fact that he was a quiet and amiable man.

I will specifically remind you of the evidence that was given by Sergeant Newman. Sergeant Newman gave evidence to the effect that he had, after Mr Melbourne was arrested, done a check of his record, his criminal record, and that there were no matters on that record other than a conviction for an offence of exceed[ing] .08 with a reading of .23 in 1975. Sergeant Newman had gone on to say that from his investigations there were no other convictions for any other matters in the Northern Territory or anywhere, and that Mr Melbourne was not adversely known to police.

The direction I am giving you is this: that when you consider that evidence as to good character, that you are entitled to consider the improbability of Roy Melbourne committing the instant offence, having a history of good character of some sixty-one years, and that this is of considerable significance.

No complaint was made by defence counsel at the trial with respect to this direction. Defence counsel say that they did not deliberately abstain from seeking a further redirection for tactical or other reasons: they simply overlooked that a different direction from the one they had sought had been given.

There is no doubt that a failure to seek a redirection is relevant in two respects [10] : as bearing upon whether on appeal the point should be allowed to be taken for the first time, and the light that the absence of an application for a redirection tends to shed on the atmosphere, and the forensic conduct of an accused's counsel, at the trial. Further, r 86.08 of the Supreme Court Rules NT [11] requires that an appellant obtain the leave of the Court of Criminal Appeal before taking a point there that was not taken at the trial.

1. See Gipp v The Queen (1998) 194 CLR 106. 2. Rule 86.08 provides as follows: "No direction, omission to direct or decision in relation to the admission or rejection of evidence of the Judge of the court of trial shall, without the leave of the Court of Criminal Appeal, be allowed as a ground for appeal, or for an application for leave to appeal, unless objection was taken at the trial to the direction, omission or decision by the party appealing or applying for leave to appeal."

I am prepared to proceed upon the basis that in this case there was no deliberate abstention from seeking a further redirection, and that the appellant should be permitted to argue that a further redirection, if sought, should and would have been given [12] : in short, that this is a case in which leave should be given if it is necessary that it be sought. Indeed it might even be arguable that the point was properly taken in any event when the application for the direction was made in the first instance.

Counsel for the appellant referred to some passages in the evidence and in the respondent's final address to the jury as demonstrating, it was submitted, that the prosecution was contending that the jury should not accept the evidence of the appellant's experts, because what the appellant had told both the police and the experts was deliberately designed to assist his defence and was untrue or otherwise unreliable.

These were the passages in the final address to which reference was made:

It's the [Lezak] test which he applies to the 15 numbers or the 15 letters test which he applies to ensure that the person is doing his best during the testing, and you might remember that was the only test missing from all his test results, the results that he takes meticulously, and my learned friend suggested keeps meticulously, this is one that he throws out and you might think that is a little bit strange, this being the one which really makes all the others genuine - and, of course, I'll come back to this later - but these tests are ones which are relied upon by Doctor Walton and Doctor Vine very much to form their own views of the brain damage, etcetera.

As to the appellant's absence of memory of the stabbing the prosecutor said:

You might remember there is something in what Mr Melbourne says the next morning indicating there are still areas that - poor memory and what might be masking or might be genuine failure to remember.

The memory impairment that we've heard described shows up on tests, not too bad when you listen to the man, see him answer questions.

You might note that it's suggested that Mr Melbourne remembers - that's how it's put as I understand it - that Mrs Chambers didn't attack him. He was asked whether she attacked him and he said "Oh no, she didn't attack me". Now, how does he know she didn't attack him if he forgets those events. That's an interesting question, isn't it. How does he know? Why doesn't he say; "I don't know"? He says "No, she didn't attack me." There's a missing link there somewhere. My learned friend might say; "Oh, he's being chivalrous again." On the other hand you might think there is some little chink there, something in his memory that indicates that to him. It's suggested that he's trying to fill in gaps. [W]ell, the gaps he fills in are pretty good, aren't they?

You might remember that Doctor Vine said, that one of the things she noticed about his memory as being a bit funny was that he hadn't remembered coming home during the day, he'd gone into town and come home. But, of course, that was in response to particular questions and when you look at the video, if you do, and you can do this more easily by going to the transcript

I'll just read this to you Mr Melbourne says; "That's why I took it home" - and he's talking about the chicken - "for us to eat, you know. I had no fridge left in the place. I said, "I may as well eat what I can of it and then chuck it away because it'll be no good in the morning." I sold my fridge in the morning. A fellow came and took my fridge away in the morning, see." Now he remembers that he's at home, he remembers the man coming for the fridge. So to that extent - that's a small point you might think - to that extent Doctor Vine, in relying upon that as an error of memory, is herself incorrect. Now, what I've suggested to you, and you'll follow this through yourself, is that his memory is pretty good for a bloke who is - who seemed to be in the problem that he was.

On the issue of the appellant's abuse of benzodiazepine the prosecutor said:

You might also think to yourself the quantity of drugs that we have proof that was taken during the relevant period is not a very high number. There were 75 Benzodiazepam - I think's the name - prescribed in the period of months prior to the offence - 75 in total. If he was taking eight or 10 at a time he wouldn't have been able to do it for many days.

There's some suggestion of him buying them around the streets and getting them from friends or trying - or something like that, you haven't got much evidence about that have you to act on, bearing in mind these issues are issues which the defence raises and the defence needs to satisfy you. So remember that onus of proof again when you get to that stage.

With respect to the appellant's level of alcoholism this was submitted by the Crown to the jury:

The other thing that you might think is significant in assessing whether Mr Melbourne is affected by alcohol to the extent that's been suggested is the way he has presented, his presentation both to his friends, his neighbours, to doctors, when he's interviewed the next day after the events of the night before, you might take those into account. You come here with knowledge of people in the world and you've seen people who drink too much and drink too much to excess and you're entitled to take into account those matters of your knowledge of the world affairs.

One of the matters that was dealt with was [the] question of peripheral neuropathy, the problem with his toes, and it is suggested by Doctor Barclay that that might have indicated an alcoholic problem of some kind and he looked at that. Doctor Welch, on the other hand, thought it was probably a problem with his feet and his treatment or suggestion was a change of shoes. We don't know which of the two solved the problem because shortly after that Mr Melbourne took himself off to Mount Isa. Nevertheless, one is as likely as the other you might think in the circumstances, and there is no continued problem, as far as we know it, and no other suggestion of this being a problem after that time when he'd got back on the drink again. So, if you look for a profile of an alcoholic person, this man doesn't fit it, I suggest to you. This is a man who had too much to drink on 27 July and on other occasions, but otherwise should not be regarded by you as an alcoholic.

Under cross-examination, Mr Taylor conceded that, when he was interviewing and testing the appellant, the appellant knew that he was "a psychologist assessing him" "for the purposes of providing evidence in court". He was cross-examined regarding his failure to keep the written Lezak test done by the appellant. He was then cross-examined about the appellant's score on the Lezak test. Mr Taylor's evidence was that the appellant scored nine out of fifteen. The Crown Prosecutor then put:

I'd suggest [to] you that twelve would be the minimum you'd expect on someone giving it a fair go?

But I'm suggesting to you Mr Taylor, nine from what you say seems to be the bare minimum, almost everybody got nine?

Q. Would it not be appropriate sir - and this is the point I make, to keep the results of that test so that they can thereafter be used and shown, demonstrated? A. Do you know I've never been asked for them before.

Q. See I suggest to you without a Lezak test and the results of it, creates some doubt in respect of the other tests as far as the patient's concerned? A. Creates doubt for whom?

185. Other examples in the cross-examination of Mr Taylor may be noted. Mr Taylor was asked if he had turned his mind to the question whether the appellant gave Mr Taylor "a true history". He was cross-examined as to the possibility that the appellant had "heard from someone" the word "paranoid" and that his use of the word involved the appellant's "justifying what he's done".

186. The appellant in his record of interview had repeatedly disclaimed any recollection of the stabbing. One particular example was as follows:

Newman: Mm-hm. I've explained to you what happened to Rene, right, and you're accused of doing that, you can't remember anything about it, can you give me any explanation as to what - what may have made you do this?

187. To support the defence case, evidence of the appellant's good character was adduced in cross-examination of prosecution witnesses. This evidence included: (a) evidence that he had no criminal record (other than a drink driving offence in 1975) and was not "adversely known to the police"; and (b) evidence that he was "amiable", "quiet", "gentle", "well behaved" and never "aggressive".

188. It may be easy to disparage evidence of this kind. Every person has a good character until he or she offends against the law. Character evidence came to be regarded as relevant following the introduction of more formal curial proceedings when local knowledge became less important and the independent role of the jury more prominent [13] . Plucknett refers to early uncertainties surrounding the admissibility of character evidence [14] :

Evidence given by witnesses to a jury was for a long time an informal adjunct to legal proceedings rather than part of their essence. It is not surprising, therefore, that there was hardly any law governing its admissibility - evidence of previous convictions, for example, was admitted without comment. 1. "An independent, original knowledge of the facts was attributed to the jury, and not a merely inferential and reasoned knowledge": Thayer, A Preliminary Treatise on Evidence at the Common Law (1898), p 137. See also Holdsworth, A History of English Law, 3rd ed (1922), vol 1, p 317: "The jury was a body of neighbours called in, either by express law, or by the consent of the parties, to decide disputed questions of fact. The decision upon questions of fact was left to them because they were already acquainted with them, or if not already so acquainted with them, because they might easily acquire the necessary knowledge. For this reason it has been said that the primitive jury were witnesses to rather than judges of the facts [T]hey represented the sense of the community - hundred or shire - from which they were drawn; and in the days when such communities had each its court, when individuals lived more simple and more similar lives, the sense of the community was a thing more distinctively realised." In Bushell's Case (1670) Vaugh 135 at 147 [ 124 ER 1006 at 1012], Vaughan CJ said: "[The jury] may have evidence from their own personal knowledge, by which they may be assured, and sometimes are, that what is deposed in Court, is absolutely false The jury may know the witnesses to be stigmatized and infamous, which may be unknown to the parties, and consequently to the Court." 2. Plucknett, A Concise History of the Common Law, 5th ed (1956), pp 436-437.

I cannot in principle make any distinction between evidence of facts, and evidence of character: the latter is equally laid before the jury as the former, as being relevant to the question of guilty or not guilty: the object of laying it before the jury is to induce them to believe, from the improbability that a person of good character should have conducted himself as alleged, that there is some mistake or misrepresentation in the evidence on the part of the prosecution, and it is strictly evidence in the case. 1. (1837) 7 Car & P 673 at 674-675 [173 ER 295 at 296].

190. The prosecution here did not adduce any evidence of bad character in rebuttal, as it might have done if it were available and the Crown wished to lead it [16] .

1. R v Waldman (1934) 24 Cr App R 204; R v Winfield (1939) 27 Cr App R 139 at 141; Stirland v Director of Public Prosecutions [1944] AC 315 at 326-327; Selvey v Director of Public Prosecutions [1970] AC 304; R v Bracewell (1978) 68 Cr App R 44; R v Hamilton (1993) 68 A Crim R 298 at 299; R v Durbin [1995] 2 Cr App R 84.

The appellant relies on all that I have quoted to found a submission that the credibility of the appellant's out-of-court assertions to the investigating police and to the expert witnesses was a matter of considerable significance and in issue at the trial: and that the absence of any issue regarding the objective facts of the killing did not mean, as the Crown claimed in a submission which Gallop J regarded as having "much force", that (in effect) questions of the appellant's credibility did not arise.

It was accordingly contended by the appellant that the Court of Criminal Appeal erred in holding that because it was the appellant's state of mind which was being examined by the experts and thus by the jury, whether the accused should be believed in terms of factual dissertations (made to the experts) was an "extraneous question", which was the language used by Gallop J.

The appellant submits that the evidence of the appellant's character was relevant in two respects: as to a normal state of non-violence making it unlikely that if his state of mind had been normal, he would have committed the crime; and as to credibility generally, even in a case in which, as here, the appellant did not give evidence.

A related submission was that the Court of Criminal Appeal erred in holding that: "evidence of good character [may not] be used to support the credibility of the accused in interrogation by police or the truthfulness of the history given to experts as a foundation of their respective opinions."

The relevance and admissibility of character evidence in cases in which an accused has not given evidence has recently been discussed in the United Kingdom.

In R v Vye [17] , Lord Taylor of Gosforth CJ (with whom Judge and Hidden JJ agreed) said:

[I]f a defendant of good character does not give evidence and has given no pre-trial answers or statements, no issue as to his credibility arises and a first limb direction is not required. 1. [1993] 1 WLR 471 at 476; [1993] 3 All ER 241 at 245.

Vye was discussed at some length in R v Aziz [18] by Lord Steyn (with whom Lord Goff of Chieveley, Lord Jauncey of Tullichettle, Lord Browne-Wilkinson and Lord Mustill agreed). His Lordship said [19] :

The certified question, although phrased in very general terms, was intended to raise the problem whether a defendant without any previous convictions may "lose" his good character by reason of other criminal behaviour. It is a question which was not directly before the Court of Appeal in Vye [20] . It is a complex problem. It is also an area in which generalisations are hazardous. Acknowledging that a wide spectrum of cases must be kept in mind, the problem can be illustrated with a commonplace example. A middle-aged man is charged with theft from his employers. He has no previous convictions. But during the trial it emerges, through cross-examination on behalf of a co-defendant, that the defendant has made dishonest claims on insurance companies over a number of years. What directions about good character, if any, must the judge give?

Counsel for the Crown and the respondents made contradictory submissions as to the correct approach. Counsel for the Crown submitted that a trial judge has a general discretion to decide whether a defendant without previous convictions has lost the right to directions in accordance with Vye by reason of other criminal behaviour. Counsel for the respondents argued that a defendant without previous convictions is always entitled to directions in accordance with Vye but that the judge is entitled to ensure that a balanced picture is placed before the jury by adding such qualifications as seems to him appropriate.

A good starting point is that a judge should never be compelled to give meaningless or absurd directions. And cases occur from time to time where a defendant, who has no previous convictions, is shown beyond doubt to have been guilty of serious criminal behaviour similar to the offence charged in the indictment. A sensible criminal justice system should not compel a judge to go through the charade of giving directions in accordance with Vye in a case where the defendant's claim to good character is spurious. I would therefore hold that a trial judge has a residual discretion to decline to give any character directions in the case of a defendant without previous convictions if the judge considers it an insult to common sense to give directions in accordance with Vye. I am reinforced in thinking that this is the right conclusion by the fact that after Vye the Court of Appeal in two separate cases ruled that such a residual discretion exists [21] .

1. R v H [1994] Crim LR 205; R v Zoppola-Barraza [1994] Crim LR 833.

That brings me to the nature of the discretion. Discretions range from the open-textured discretionary powers to narrowly circumscribed discretionary powers. The residual discretion of a trial judge to dispense with character directions in respect of a defendant of good character is of the more limited variety. Prima facie the directions must be given. And the judge will often be able to place a fair and balanced picture before the jury by giving directions in accordance with Vye [22] and then adding words of qualification concerning other proved or possible criminal conduct of the defendant which emerged during the trial. On the other hand, if it would make no sense to give character directions in accordance with Vye, the judge may in his discretion dispense with them.

Subject to these views, I do not believe that it is desirable to generalise about this essentially practical subject which must be left to the good sense of trial judges. It is worth adding, however, that whenever a trial judge proposes to give a direction, which is not likely to be anticipated by counsel, the judge should follow the commendable practice of inviting submissions on his proposed directions.

1. [1996] AC 41 at 50-51. 2. R v Aziz [1996] AC 41 at 52-53. 3. [1993] 1 WLR 471; [1993] 3 All ER 241. 4. R v H [1994] Crim LR 205; R v Zoppola-Barraza [1994] Crim LR 833. 5. [1993] 1 WLR 471; [1993] 3 All ER 241.

I would respectfully agree generally with these observations. Prima facie the direction should be given. Ordinarily strong contra-indicative factors would have to be present before a trial judge should conclude that such a direction is not to be given. The fact that an accused has not given evidence is not of itself a conclusive reason for declining to give the direction.

That evidence of good character may be used to support the credibility of statements made out of court by accused persons has been accepted in New Zealand [23] . The appropriateness of its use was also endorsed in New South Wales in R v Gillard [24] and in South Australia in R v Trimboli [25] . In the former case [26] Gleeson CJ said:

It is well established that character evidence may be used on the issue of the appellant's credibility [27] . As was pointed out in Trimboli [28] , character evidence supports the credibility of the appellant's account "as to the objective facts and leav[es] more room for acceptance of his evidence of his own subjective state of mind" [29] . Here the information given by the appellant to medical practitioners was said to be false or exaggerated, and in that respect his credibility was at stake. 1. R v Falealili [1996] 3 NZLR 664 at 666-667. 2. Unreported; Court of Criminal Appeal; 15 July 1991. 3. (1979) 21 SASR 577. 4. R v Gillard (unreported; Court of Criminal Appeal; 15 July 1991), at 10. 5. Attwood v The Queen (1960) 102 CLR 353; R v Murphy (1985) 4 NSWLR 42 at 54. 6. (1979) 21 SASR 577 at 586-588. 7. Trimboli (1979) 21 SASR 577 at 588.

In both the Commonwealth and New South Wales, the Evidence Act 1995 contemplates the possibility of the use of a good character direction to support the credibility of an accused's statements out of court. Speaking of the New South Wales Crimes Act 1900, s 412, a Court of Appeal of five judges (Street CJ, Hope, Glass, Samuels and Priestley JJA) said this [30] :

[W]hilst the primary significance of evidence of good character is upon the unlikelihood of guilt, there is a corollary to the effect that evidence of good character can be used with reference to credibility of the accused in his denial of the charge, and hence the unlikelihood of his guilt. The omission to give a specific direction on the credibility aspect may or may not be regarded as resulting in a miscarriage, according to the particular circumstances of the case in hand. 1. R v Murphy (1985) 4 NSWLR 42 at 54.

I have formed the view that the appellant has made out a case that his reliability and credibility were issues at the trial notwithstanding that he did not give evidence and that the substantial issue at the trial was whether the appellant was in a state of diminished responsibility at the time of the stabbing. It seems to me that his credibility was at stake in the sense suggested by Gleeson CJ in Gillard in the passage that I have set out. It does appear to me that the Crown Prosecutor, guardedly but nonetheless plainly, and in no way improperly, in his cross-examination of the experts, did seek to put in issue not just the reliability of the appellant but also his credibility.

A question was raised about the relevance of the appellant's out-of-court assertions and the use to which they could be put. The statements by the experts which I have quoted, repeating out-of-court assertions by the appellant and statements by him recorded in the record of interview, were received as original evidence. No objection was taken, or any qualification sought to be made to them at the trial. The record of interview was tendered and relied on by the Crown. The jury were entitled to believe or not believe the statements that were made out of court in that record. That they might do so meant that they necessarily had to give consideration to the appellant's honesty. That in turn meant inevitably that his character for this purpose was a relevant factor in determining his guilt.

There is a distinction that may on occasions be made in relation to psychiatric evidence from other medical evidence. Sometimes in the case of the former, what the patient has said, and is saying may be a manifestation of the illness if it exists, or an indication that it does not exist, in the same way as, for example, the presence or absence of a discolouration of the skin may be a manifestation of a physical illness such as jaundice. In each case what can be heard or seen is a symptom upon which reliance will and may properly be placed for the diagnosis. Bizarre assertions may be a form of bizarre conduct: the assertions are the conduct. Once it has been proved that the assertions have been made, the fact of the making of those assertions and their content are matters that a psychiatrist may take into account in forming his or her opinion. Evidence of statements of this kind may therefore be received as original evidence. It is true that a jury will not have the same opportunity of assessing such evidence as in the case of other original evidence consisting of statements made in court but a jury is in this respect in no different a position from what its members would be in when a doctor describes in evidence symptoms of a patient after the illness is cured or those symptoms have disappeared. It will also probably be the position that if the accused in such a case does not give evidence to enable the jury to form their own opinion of him or her, the "statements", being conduct as recounted by others will be less persuasive than they might otherwise be, but that does not mean that in an appropriate case involving mental infirmity the evidence of the statements to the extent that they have been relied on by the experts should not be admitted.

Since opinion evidence involves the drawing of inferences and conclusions from facts, the admissibility of such evidence depends upon proof or admission of the facts upon which the opinion is based [32] . 1. (1990) 49 A Crim R 243 at 249. 2. Murphy v The Queen (1989) 167 CLR 94 at 120; Ramsay v Watson (1961) 108 CLR 642; Paric v John Holland (Constructions) Pty Ltd (1985) 59 ALJR 844 at 846; 62 ALR 85 at 87-88; R v Turner [1975] QB 834 at 840; Harmony Shipping Co SA v Saudi Europe Line Ltd [1979] 1 WLR 1380; [1979] 3 All ER 177.

The facts upon which psychiatric or psychological opinions are based may include the fact that a person has made statements in the form, to the persons and on the occasions upon which they were made. If they were not made genuinely, in the sense of, not honestly made, then the opinions of the doctors based on them would be shaken or indeed perhaps even demolished. The genuineness, that is, the absence of simulation in the making of the statement may therefore be a matter of relevance at a trial, and an issue upon which an accused person's good character could have a bearing.

In this case however very little of what the appellant told the doctors and the psychologist falls into the category of evidence which I have been describing. Most of it consisted of assertions by him to them out of court of the existence of various addictions, absence of memory, beliefs held by him, and claims of certain personality traits, and could not therefore be regarded as manifestations of abnormality by conduct. Ordinarily evidence of assertions of that kind could and should only be received if the maker of the assertions also gives them in evidence at the trial, and should only be referred to in advance by other witnesses if the accused is to give evidence of them at the trial. It is only because of the absence of objection to them in this case and the way in which the trial was conducted generally that regard may be had here to the assertions and the experts' reliance on them.

The evidence however of what the appellant said to other witnesses shortly after the event is capable of falling into the category of evidence of conduct and was admissible as original evidence going to the appellant's state of mind at the time. The experts were, therefore, entitled to rely on it as evidence of the appellant's state of mind, and a jury could give both it and the experts' opinion on it such weight as they saw fit.

Ramsay v Watson [33] and R v Schafferius [34] were relied on by the respondent. What was said in the former is not, with respect, to be doubted:

This makes all statements made to an expert witness admissible if they are the foundation, or part of the foundation, of the expert opinion to which he testifies; but, except they be admissible under the first rule, such statements are not evidence of the existence in fact of past sensations, experiences and symptoms of the patient. Hearsay evidence does not become admissible to prove facts because the person who proposes to give it is a physician. And, if the man whom the physician examined refuses to confirm in the witness box what he said in the consulting room, then the physician's opinion may have little or no value, for part of the basis of it has gone. Each case depends on its own facts. 1. (1961) 108 CLR 642 at 649. 2. [1977] Qd R 213.

The Court was not there however considering a case of mental disability or as here a case in which the evidence was received without objection and discussed at length at the trial by both parties.

In R v Schafferius [35] , although the issue was one of diminished responsibility, the evidence that the psychiatrist gave of the statements made to him by the accused was not evidence which could in any way be regarded as "conduct" in the sense that I have discussed it, or as a foundation for any retrial on it for that purpose. The evidence in question there was that the accused "had a deep love" for the young woman he had killed.

In my opinion the credibility of the appellant was in issue in this case although he did not give evidence. That issue arose in respect of some evidence which was admissible and other evidence which although inadmissible if objected to, was not the subject of any objection. A credibility direction should therefore have been given. This is not an appropriate case for the application of the proviso as I cannot say that the appellant has not lost a real chance of a verdict of diminished responsibility [36] . I would allow the appeal and order a retrial.

1. See Mraz v The Queen (1955) 93 CLR 493; M v The Queen (1994) 181 CLR 487; BRS v The Queen (1997) 191 CLR 275.