Tarrant v R

1. [1]

   THE COURT: For some nine years prior to 29 July 2013 Alois Rez had lived in a de facto relationship of some turbulence with Sarah Tarrant. On the night of 29 July 2013 he was killed. The Crown alleged that he was murdered by Ms Tarrant and a man with whom she was then in a sexual relationship, Raymond Roff. Ms Tarrant did not deny her involvement in the killing of Rez and proffered a plea of guilty to a charge of manslaughter, which was rejected by the prosecution. Accordingly, she went to trial in March 2016 on a charge of murder, before Fagan J and a jury. On 14 April 2016 the jury brought in a verdict of manslaughter on the basis that she suffered from a substantial impairment of her mental capacities for the purposes of s 23A of the Crimes Act 1900 (NSW).

2. [2]

   Immediately following her conviction, Mr Roff’s trial commenced, also before Fagan J and a jury. Ms Tarrant was the principal witness against Mr Roff, who denied any involvement in the killing. Her evidence, consistently with her evidence at her own trial, was that he had carried out the killing after she had administered sleeping tablets to Mr Rez in his evening meal. On 13 May 2016 the jury found Mr Roff guilty of the murder of Mr Rez.

3. [3]

   The present applicant was not sentenced until 19 August 2016. On that date, the trial judge imposed a sentence of imprisonment of 10 years and 8 months, with a non-parole period of 8 years. On the same day, Mr Roff was sentenced to imprisonment for 32 years with a non-parole period of 24 years. The applicant now seeks leave to appeal against her sentence. An appeal by Mr Roff with respect to his sentence was upheld by this Court, which reduced his sentence from 32 years to 25 years, with a non-parole period of 18 years 9 months. [1]

Issues on appeal

1. [4]

   Unsurprisingly, the variation in Mr Roff’s sentence was relied upon by the present applicant as a basis for alleging a lack of proportionality between the respective sentences. [2] Otherwise, the matters sought to be raised by way of appeal focused on two broad issues. First, the applicant sought to challenge the refusal of the sentencing judge to disqualify himself in response to an application by counsel for the applicant dated 24 June 2016. The court had convened on the afternoon of 24 June 2016 to deal with the sentencing of Mr Roff. However, at the outset of the hearing, the judge noted that he had received a written submission from senior counsel for Ms Tarrant seeking his recusal with respect to her sentencing. He stated that he would not accede to the application, indicating that he would give his reasons later. [3] The reasons were published four days later on 28 June 2016. [4]

2. [5]

   The second matter relied on by the applicant, which was not confined to a single ground, concerned the manner in which the sentencing judge assessed the substantial impairment of the applicant’s mental capacity, which formed the basis of the jury’s verdict of manslaughter. It had been common ground between the prosecution and the defence at trial that the applicant suffered from a substantial impairment; the question left to the jury was whether the impairment was “so substantial as to warrant liability for murder being reduced to manslaughter.” [5] The applicant submitted that, in assessing the relative culpability of the applicant on sentencing, the trial judge failed to give full weight to the verdict of the jury with respect to that issue.

3. [6]

   The gravamen of the principal substantive ground was that the sentencing judge had diminished the force and effect of the unchallenged psychiatric evidence. He did that, it was submitted, by dissecting the conduct of the applicant upon which the opinions were based and reassessing that conduct as of diminished significance, or rejecting the conduct as not established on the evidence, even where the evidence was unequivocal and unchallenged.

4. [7]

   At least where the appeal is not one limited to a question of law, [6] the established principle is that a ground challenging the authority of the court to determine a matter, including a submission concerning apprehended bias, should be addressed before turning to other grounds of appeal. Thus, in Concrete Pty Ltd v Parramatta Design & Developments Pty Ltd [7] Gummow ACJ held that the Full Court of the Federal Court had fallen into error by dealing first with the substantive issues and then with the alleged appearance of bias on the part of the trial judge, stating: [8]

5. [8]

   The principle has been accepted and applied in numerous cases. [11] Senior counsel for the applicant accepted (indeed, insisted) that if she were to succeed on the apprehended bias ground, the sentence should be set aside and the matter remitted to the Common Law Division for resentencing. In effect, the power of this Court under s 6(3) of the Criminal Appeal Act 1912 (NSW) to substitute another sentence would not be engaged in circumstances where no valid sentence had yet been determined. Senior counsel for the Director of Public Prosecutions did not take issue with that approach. Accordingly it is necessary to address first the material said to establish prejudgment.

(a) legal principles

1. [9]

   There was no issue in this Court that, where actual bias is not alleged, the legal test, as expressed in Johnson v Johnson, requires the court to be satisfied that “a fair-minded lay observer might reasonably apprehend that the judge might not bring an impartial and unprejudiced mind to the resolution of the question the judge is required to decide.” [12] The High Court has reaffirmed the salience of the test on subsequent occasions, including in Michael Wilson & Partners Ltd v Nicholls. [13] The so-called “double might” test is by no means easy to apply: its application requires attention to four discrete elements.

   1. (1)

      First, there is the postulate of the “fair-minded lay observer”. Use of that expression reflects the principle that it is public confidence in the administration of justice which is sought to be preserved and not, as explained in Johnson v Johnson, [14] “the assessment by some judges of the capacity or performance of their colleagues.” [15]

   2. (2)

      Secondly, the test has been described as “objective”, by which is meant a third party’s assessment of the judge’s conduct and capacity, and not, as with actual bias, an assessment of the judge’s own state of mind.

   3. (3)

      Thirdly, there is said to be a two-stage process required; it is necessary to articulate “the connection between the events giving rise to the apprehension of bias through prejudgment and the possibility of departure from impartial decision-making.” [16]

   4. (4)

      Fourthly, use of the term “might” lowers the burden of proof below that of probabilities. Thus the court need not be satisfied that the fair-minded lay observer “would” have such an apprehension; nor need any apprehension in the mind of the observer itself involve a state of satisfaction on the probabilities.

2. [10]

   In the present case, there was an issue as to what level of understanding should be attributed to the lay observer with respect to the role of the judge in a criminal trial. The purpose of protecting confidence in the administration of justice is itself in danger of being undermined if the level of sophistication required of the lay observer is overstated. Some factors will inevitably be indeterminate and will therefore require evaluative judgment. In British American Tobacco Australia Services Ltd v Laurie [17] , the applicant had submitted that “to attribute knowledge of the Tribunal’s statute to the lay observer is to endow that hypothetical construct with a degree of legal knowledge that is likely to be enjoyed only by practitioners who appear regularly before the Tribunal.” [18] In response, the joint reasons of the majority said, in somewhat general terms, referring to the earlier discussion in Johnson v Johnson: [19]

3. [11]

   Rouvinetis v Knoll held: [20]

4. [12]

   There is also a level of unavoidable imprecision in the standard of what a person “might” apprehend. Clearly a fanciful or speculative possibility must be put to one side and the reasonable fear of the observer must be “firmly established”. [21]

5. [13]

   Finally it is necessary to identify the material to which the Court may have regard. Where the recusal application is based on the conduct of the judge during the various hearings in open court, the court may, of course, have regard to what was said and the context in which it was said. The source of concern may be clarified by later statements, which should, therefore, not be ignored.

(b) recusal judgment

1. [14]

   That approach raises a question as to the judge’s response to the recusal application. The cases treat material which is not itself the cause of the apprehension in discrete ways. First, the impugned statement must be read in context and must take account of subsequent statements which may be sufficient to eradicate any reasonable apprehension of bias. [22] In Johnson v Johnson, the joint reasons stated: [23]

2. [15]

   Secondly, may the reasons given for refusing a recusal application also be relied upon? At least in principle, reasons given for rejecting a recusal application may constitute a later clarification. They may reveal information which the hypothetical bystander would accept as demonstrating that an earlier opinion was based on a misunderstanding. Alternatively, the reasons may provide support for the earlier opinion. However, in the latter case, there is a risk that a recusal judgment may tend to demonstrate actual bias, with its focus on the subjective state of mind of the judicial officer. That is one reason why the High Court has warned against reliance on the final reasons for judgment in the matter. [24]

3. [16]

   It is clear that reasons for the final judgment fall into a different category. Not only is there a risk they will be relied on to demonstrate actual bias, they also carry the risk of inferring bias from an adverse outcome. There may, of course, be legitimate reasons for looking at the final judgment. For example, it may disclose for the first time that the judge had an inappropriate interest in the proceedings, or undertook inappropriate consultations; or it may identify the issues in dispute. Appropriately, the applicant did not rely on the sentencing judgment under this ground.

4. [17]

   The recusal judgment is of limited assistance, for three reasons. First, many of the complaints concerned what were said to be inappropriate judicial interventions in one or other of the trials. This distinction was not consistently recognised in the recusal application. Further, even if those complaints had merit, they did not, as such, provide a basis for a reasonable apprehension of bias. Some judges are more interventionist than others; to demonstrate an apprehension of bias, it is necessary to have regard to the content of the interventions and whether they were generally adverse to one party.

5. [18]

   Consequently, much of the recusal judgment was devoted to a demonstration that the interventions were consistent with a proper exercise of the judicial function in the course of a jury trial. Minds may differ as to the degree to which it is desirable for a judge to question witnesses in the course of their examination in chief or their cross-examination when both parties are represented by experienced counsel. The jury is required to determine factual issues without regard to the views of the trial judge. Although such complaints may need to be determined where raised on appeal against a verdict, that did not happen in the present case; Ms Tarrant was convicted only of the offence to which she had proffered a plea and, although Mr Roff was convicted of murder, she was a prosecution witness at his trial so that any hostility towards her on the part of the trial judge could not form the basis of an appeal by him. There being no appeal, it is not necessary to resolve these complaints.

6. [19]

   The second reason for placing little weight on the recusal judgment is, in general terms, explained by the joint reasons in Laurie: [25]

7. [20]

   Thirdly, the recusal judgment is of limited assistance because it is by no means clear that the judge applied the correct test. It is true that he identified the correct test by reference to the relevant passages in Ebner v Official Trustee in Bankruptcy [26] and Johnson v Johnson. [27] The question is, rather, whether the judge applied the test when addressing the aspects of his conduct relied upon by the applicant to demonstrate a reasonable apprehension of bias. However, this is not an appeal from the recusal judgment. Failure to apply the correct test would not demonstrate that the recusal application should necessarily be upheld. Accordingly, it is sufficient to give an example to explain the basis on which the recusal judgment should be put to one side. It will then be necessary for this Court to form its own view as to the ground of appeal based on a reasonable apprehension of bias, which in fact raised issues going beyond the original recusal application to include events which occurred thereafter.

8. [21]

   The recusal application referred to a series of six questions asked by the judge which appeared to express disbelief in her evidence that she believed Mr Roff’s ultimatum that he would leave Dubbo if Rez were not “gone” within three weeks. [28] One particular intervention occurred during the cross-examination of the applicant by counsel for Mr Roff (Mr Sutherland SC) in the course of his trial. [29]

9. [22]

   The judge sought to justify this intervention on the basis that he “did not consider [Ms Tarrant] was making a conscientious endeavour to answer counsel’s question honestly and to the best of her ability.” [30] The judge continued: [31]

10. [23]

    It may be accepted that rejecting the proposition that “a reasonable observer properly informed could entertain a reasonable apprehension” involves the rejection of the affirmative proposition that the reasonable observer “might” hold such an apprehension. In other passages, the judge suggested that the reasonable observer “could not” hold such an apprehension, which again appears to negative the possibility that such an apprehension might be held, although the consistent departure from the language in which the test is expressed is undesirable. On the other hand, that which could not be entertained was “that that judge would hold a bias against the witness ….” That language applies too high a standard to the objector. The test of the observer’s state of mind is not whether he or she might believe that the judge “would” be biased, but whether the judge “might” be biased.

11. [24]

    It is not necessary to give further examples; the judgment dealt clearly and comprehensively with the numerous allegations raised by the recusal application and was prepared and delivered within four days, in circumstances where the judge anticipated the possibility of an interlocutory appeal before the next hearing date and did not wish the absence of his reasons to lead the sentencing hearing to be adjourned.

12. [25]

    It is necessary then to turn to the first ground of the current appeal, which primarily reiterated the claims of a reasonable apprehension of bias raised in the recusal application, together with later expressions of opinion by the trial judge during and after the Roff trial.

(c) applicant’s case on appeal

1. [26]

   Before addressing the elements of ground 1 (alleging apprehended prejudgment) it is helpful to note, by way of context, the basis of the substantive grounds. In broad outline, the applicant’s case on appeal was as follows:

2. [27]

   While the judge accepted that he was bound to sentence on the basis that the jury verdict accepted the fundamental elements of the offence, because the jury had relied upon the opinions of the psychiatrists, he considered it was necessary for him to be satisfied that the factual assumptions underlying those opinions had been made good, and the links between facts and opinions exposed. The applicant accepted that in considering her culpability, the judge was required to determine how severely her freedom of choice was constrained by the history of domestic abuse and her mental conditions. However, she submitted that that exercise did not permit a re-evaluation of the circumstances underlying the expert opinions.

3. [28]

   These were legitimate grounds of appeal and would, if established, require this Court to intervene and reconsider the applicant’s sentence. However, the adoption of an erroneous approach to the sentencing exercise does not of itself demonstrate prejudgment. Yet some of the factual elements said to give rise to a reasonable apprehension of prejudgment amounted to no more than an articulation, at different stages and in different words, of the substantive grounds.

4. [29]

   The applicant’s written submissions identified several bases upon which the judge was said to have demonstrated prejudgment, namely:

   1. (1)

      comments immediately following her conviction on 14 April 2016 –

   2. (2)

      an indication on 18 April 2016 that, if Mr Roff were convicted, the judge would conduct joint proceedings with respect to sentencing;

   3. (3)

      aspects of his questioning of the applicant during the Roff trial;

   4. (4)

      questioning of the applicant during her own trial as to why she did not seek a second apprehended violence order, and

   5. (5)

      further remarks as to the basis of the expert opinions during a directions hearing on 13 May 2016 in the absence of the applicant, and

   6. (6)

      questioning of the experts at a sentencing hearing.

5. [30]

   It is convenient to deal with the particulars of prejudgment in the order set out above.

(d) comments made on 14 April 2016

1. [31]

   Two matters arose immediately following the return of the verdict on 14 April 2016. The first concerned the question of bail pending sentence, the applicant having been granted bail prior to the trial and remained at large during the course of the trial. The judge expressed the view that he found it “very difficult to see how anything less than fulltime custody could be imposed” [33] and, in substance, invited the prosecutor to make a detention application.

2. [32]

   It was true that the applicant was in no different position following the verdict of the jury than she had been for some time, having first indicated her intention to plead guilty to manslaughter and having entered the plea at the commencement of her trial. She could have remained at large until sentenced. At that point, it was envisaged that the sentencing hearing would take place in eight days’ time.

3. [33]

   Although the submissions focused in part on what was said to be a misreading of the Bail Act 2013 (NSW) in dealing with the detention application, there is no basis for contending that that gave rise to an apprehension of bias. That apprehension could only have arisen from the fact that the judge took the initiative in suggesting that it might be appropriate to revoke bail following the verdict of the jury and that he indicated, prior to hearing submissions from either party on sentence, that he was not inclined to impose a non-custodial sentence.

4. [34]

   The judge stated that he had already been provided with cases in which non-custodial sentences had been imposed for manslaughter and considered them very different. There was no suggestion that he would not have fairly considered any submissions made on sentence by senior counsel for the applicant, at the appropriate time. So much would have been understood by the lay observer in the court; there was no realistic possibility that he or she would view the matter differently. This matter could not give rise to a reasonable apprehension of bias.

5. [35]

   The second matter, raised in the course of the hearing on 14 April, related to the use of the psychiatrists’ reports as a basis for finding substantial impairment warranting a reduction of the offending from murder to manslaughter. In substance, both psychiatrists had accepted substantial impairment of the applicant’s mental state, and the judge had directed the jury that their opinions extended to all three elements of capacity referred to in s 23A(1).

6. [36]

   When considering the likely course of the sentencing hearing, the judge remarked that: [34]

7. [37]

   Counsel for the applicant noted that the judge was “bound by the jury’s verdict” and suggested that the jury had obviously placed considerable weight on the reports. The judge continued: [35]

8. [38]

   It was not in doubt that the judge was bound by the verdict with respect to the essential elements of the charge. Nor was it in doubt that the judge was required, in sentencing the applicant, to consider the level of her culpability, as a case of manslaughter resulting from impaired mental capacity. Nevertheless, the applicant submitted that the language used by the judge might have indicated to the lay observer that he might conduct the sentencing without in truth accepting the jury’s verdict; it was not possible to accept the verdict without accepting the critical evidence upon which the jury must have relied. The effect of the submission was that the lay observer might think that the judge might not give effect to the verdict, and thus sentence the applicant as if convicted of murder.

9. [39]

   Taken in isolation, the comments set out above do not satisfy the test of apprehended bias. First, the judge was not making a finding, but commenting, albeit in strong terms, on how he should approach the sentencing task which, on any view, was not without difficulty. Secondly, he identified a specific difficulty with reconciling parts of the evidence. Although expressed in terms of exasperation or frustration, it is the kind of interchange between judge and counsel which is helpful in giving counsel an opportunity to know what topics to address, either by evidence or in submissions. Thirdly, when the very same reports were left to the jury, the judge had summarised their contents in entirely appropriate terms and, indeed, favourably to the applicant, whose counsel, at a different point in the submissions on the appeal, sought to rely expressly on the way in which the material had been left to the jury. Knowing all that, it is not plausible that the lay observer might think the judge might not respect the jury verdict or deal with the sentencing exercise otherwise than on the evidence then before the court.

(e) comments made on 18 April 2016

1. [40]

   Following the hearing on 14 April 2016, it appears that the judge sent an email to counsel and relisted the matter on the following Monday, 18 April 2016. The email is not before this Court, but two issues raised in it were addressed on 18 April, namely (a) a concern about the sentencing taking place whilst Mr Roff’s trial was underway, and (b) the appropriateness of a joint sentence hearing, if he were to be convicted. In the course of the discussions which followed (counsel for the applicant robustly disputing both suggestions raised by the judge) the judge stated: [36]

2. [41]

   In substance, Ms Manuell SC (for the applicant) submitted that there had been separate trials, the evidence might well prove to be different in each and, whatever was said at Mr Roff’s trial, the applicant would not be in a position to challenge it. As a result, the sentencing of each might need to occur on “different factual bases”. [37] The judge responded:

3. [42]

   The only result of the brief hearing on that morning was to vacate the hearing date for the sentencing which had been fixed for that Friday.

4. [43]

   In the course of the appeal, the judge’s observations were deployed as demonstrating some form of prejudgment (the precise nature of which was not clearly articulated) adverse to the applicant. As the submissions correctly noted, there was nothing ridiculous or absurd in sentencing two co-offenders on the basis of different facts where there had been separate trials. The submissions, forcefully made before the sentencing judge, to the effect that evidence given in the Roff trial should not be relied upon in sentencing the applicant, had considerable force and would need to be addressed in considering whether the sentencing exercise miscarried. Inevitably, given that no one then knew whether Mr Roff would be convicted, the issue as to the course of the sentencing proceedings was not then resolved.

5. [44]

   The only relevance of the observations made on this occasion in relation to the allegation of a reasonable apprehension of bias was that the trial judge could have been said to be following an erroneous approach which would, if executed, result in unfairness to the applicant. There was no basis to find a reasonable apprehension of bias on this material.

(f) questioning the applicant at Mr Roff’s trial

1. [45]

   The judge asked questions of the applicant during the trial of Mr Roff, at which she gave evidence central to the prosecution case. Again, the context limits the effect on any possibility of prejudgment. It is convenient to say why that is so before turning to the specific passages in the questioning now relied on by the applicant. Four contextual considerations are important. First, the questioning took place in the course of a trial of Mr Roff before a jury. The judge was not required to make any findings of fact, nor did he do so. Secondly, although the questioning was said to be unfair to the applicant because she was unrepresented at the trial of Mr Roff, that complaint is beside the point; the issue is whether the questioning revealed an element of prejudgment on the part of the trial judge. Thirdly, at the time questioning took place, the lay observer would not know (nor would the lawyers) whether evidence given by Ms Tarrant at Mr Roff’s trial would in fact be admitted on her sentencing hearing. Fourthly, it was said that the questioning was “inappropriate” in the context of Mr Roff’s trial. For example, to the extent that there might be issues as to the credibility of Ms Tarrant’s evidence, such issues should not have been canvassed by the judge in the course of her examination-in-chief, but should have been left to senior counsel for Mr Roff to deal with in cross-examination. That submission may have had some force, but it did not go to the heart of the issue. The real issue had two strands. One was the inference that the judge held such a level of scepticism as to Ms Tarrant’s truthfulness that the lay observer might fear that he might not be able to assess her evidence at the sentencing hearing with impartiality. The other strand lies in the possibility that some of the questions raised issues which were irrelevant in Mr Roff’s trial, but could have been relevant to the sentencing of Ms Tarrant. If the lay observer were to draw that inference, it would have been directly relevant to a fear of prejudgment.

2. [46]

   With those considerations in mind, it is convenient to refer to the passages relied upon by the applicant on the appeal.

3. [47]

   The first passage occurred in the course of Ms Tarrant’s evidence in relation to a letter sent to Mr Roff. She referred to the first occasion on which they had sexual intercourse and expressed regret that she had taken a “morning after” pill so that she would not become pregnant. The judge interposed to ask the following questions: [40]

4. [48]

   The applicant submitted that the purpose of the cross-examination was to show “she had capacity for moral judgment”, an issue which was raised again at the sentencing hearing. However, the issue as to capacity to make a moral judgment arose from the terms of the letter; the judge’s questions really took the matter no further.

5. [49]

   The second passage occurred when the applicant gave evidence (both in her trial and in the Roff trial) of an occasion when she had been assaulted on 13 July 2013, about two weeks before Mr Rez was killed. She and Mr Roff had been watching television in the lounge room, while the deceased was upstairs in the bedroom. Later he had sent a text message on her phone saying that she had “about half an hour to pack my stuff and get out of the house.” [41] She said that she went to talk to Mr Rez who was “very angry.” [42] An argument ensued in the hallway; the argument proceeded for about five or 10 minutes and then Rez “tried pushing me out the front, he pushed me down the hallway and tried pushing me out the front door and he punched a hole in one of the walls in the hallway.” [43] The argument continued until she went to the toilet. [44] She sent text messages about the incident to Mr Roff who called the police. They arrived “a couple of hours” later. [45] She gave evidence that things settled down after the outburst. The judge interrupted the examination with the following questions: [46]

6. [50]

   She gave evidence that she did not tell the police about the assaults because she was “too scared to tell them.” [47] The judge then intervened again:

7. [51]

   For a trial judge to cross-examine a prosecution witness, in an argumentative fashion, in the course of her evidence-in-chief is unusual and is fraught with difficulties. What the jury would do with a question in leading form including five propositions to which the witness answered “yes” is by no means clear. However, the issue for present purposes is whether, as the applicant submits, the questions demonstrated prejudgment based on a lack of understanding of the dangers faced by a physically much smaller women in the home threatened by a larger and potentially intimidating male partner, where there had been a history of abuse and contemptuous behaviour, prior violence and when the woman was pregnant. Taken in isolation, this questioning may not have demonstrated a significant level of disbelief of the applicant’s evidence, but it cannot be taken in isolation. There were further questions asked in similar vein and in part the answers were picked up and relied upon at a later pre-sentence hearing.

8. [52]

   A third example, to which it is necessary to return, was that identified earlier in dealing with the recusal judgment. Counsel for Mr Roff had cross-examined Ms Tarrant as to the development of her relationship with Mr Roff since she arrived in Dubbo in the middle of 2012. She had agreed that she started to develop some feelings towards him which she accepted were both romantic and sexual. [48] She agreed that she had sent sexually explicit photographs of herself, that they were sexually provocative and extremely so. [49] She agreed they had intercourse in the back seat of a car on the side of the road. The questioning continued: [50]

9. [53]

   The evidence as to which the judge appeared incredulous was the subject of further questioning of the psychiatrists at the sentencing hearing. Whether it involved any element of prejudgment may depend on whether the judge was entitled to form a view of the facts in the course of Mr Roff’s trial, or whether that evidence was irrelevant at the applicant’s sentencing. It will be necessary to return to this issue.

10. [54]

    The fourth passage occurred after the applicant had been cross-examined as to her discovery of Facebook messages between Rez and another woman, Amanda Vine, suggesting that Rez was planning to leave the applicant, take their children and marry Vine. His Honour then intervened: [51]

11. [55]

    Fifthly, the applicant was also cross-examined about her evidence that Roff had said to her that “[i]f Alois is not gone within three weeks” he would be leaving. [52] The judge then asked: [53]

12. [56]

    A lay observer hearing this questioning might well have formed the view that the judge did not accept her evidence on this particular topic. Again, taken in isolation it is difficult to demonstrate a connection between that questioning and the likelihood that the judge would not deal impartially with evidence she gave at a sentencing hearing, or in assessing the evidence she gave at her own earlier trial.

(g) failure to seek a second AVO or leave Rez

1. [57]

   In the course of her evidence-in-chief as the accused in her own trial, the trial judge questioned her on a number of occasions. For the most part, although reasonably extensive, the questioning was neutral. On other occasions it was argumentative. The applicant had given evidence that she obtained an apprehended violence order (AVO) in 2008. One of the terms of the order required Mr Rez not to go within 50 metres of the premises at which Ms Tarrant was living at the time, on the central coast; nor was he to approach or contact her. After a period he persuaded her to change the terms so that he could return home. The AVO nevertheless remained in place, requiring him not to assault, molest, harass, threaten or otherwise interfere with her. The following exchange took place: [54]

2. [58]

   The substance of that evidence elicited by the judge was considered in a later hearing; although it was a lengthy intervention in the leading of evidence from an accused person, no objection was taken and it does not in its terms provide any evidence of prejudgment.

(h) remarks made on 13 May 2016

1. [59]

   On 13 May 2016 the jury in the Roff trial returned a verdict of guilty of murder. There followed a discussion with counsel as to a convenient date to conduct sentencing proceedings and as to whether there could be a joint hearing of the sentencing for both Ms Tarrant and Mr Roff. No representative of Ms Tarrant was there on that occasion. At the end of the day, the matter was merely adjourned to Monday, 23 May 2016.

2. [60]

   In the course of discussion, the judge reiterated views he had expressed earlier about the desirability of a joint hearing: [55]

3. [61]

   The last sentence carried two troubling inferences, namely that (a) the jury’s verdict had been based on a false evidential premise and (b) the prosecutor had connived in this outcome by not challenging at least the basis of the psychiatric opinions.

4. [62]

   In our view, the lay observer who had followed the proceedings thus far would have been troubled as to the approach the judge was taking with respect to Ms Tarrant. First, such a person would appreciate that the judge himself had made no findings of fact relevant to the sentencing at any stage. Further, the judge had described his remarks, on more than one occasion, as tentative and put forward argumentatively. On the other hand, and despite the disclaimers, some opinions entirely adverse to Ms Tarrant were expressed in strong language; they were not expressed in tentative terms.

5. [63]

   Secondly, the observer would know that the opinions were not expressed in a vacuum, but were responses to the evidence which the judge had already heard, including from Ms Tarrant who had been in the witness box for some seven days. The observer would appreciate that there might well be further evidence at a sentencing hearing, but it appeared that the judge was expecting to take account of all the evidence he had heard at the trials. In this sense, some tentative judgments might properly be formed by reference to that evidence.

6. [64]

   Thirdly, the observer would appreciate that the thrust of many of the comments concerned the need for further evidence, in terms which strongly suggested that it would be in the interests of all parties to consider calling further evidence or even giving further evidence themselves, in the case of the offenders. In relation to Ms Tarrant, there was a clear inference that, absent further explanation, the judge might draw damning inferences against her.

7. [65]

   Fourthly, the lay observer might be deeply troubled by the fact that the last round of observations had been made in the absence of Ms Tarrant and her counsel. It is one thing to engage in discussion with counsel as to one’s tentative views, the observer might reason, but it is quite another to express adverse views about one party in their absence and in strident terms.

8. [66]

   Fifthly, the whole purpose of the principle relating to reasonable apprehension of bias (and indeed actual bias) is that the public must have confidence that the judge will make findings solely on the basis of the evidence before the court and will assess that evidence impartially. There is a logical corollary which the lay observer might be assumed to infer, namely that Ms Tarrant would be sentenced on the basis of evidence which was given in her trial, without reference to extraneous material which may have arisen in other proceedings to which she was not a party. There was an ambiguity about the repeated statements of the trial judge that he proposed to conduct a joint sentencing hearing. It was apparent that the judge envisaged that further evidence called at a joint hearing might be evidence in both proceedings, although that was not resolved. It was also not clear whether the judge would permit evidence given in one trial to be evidence in the other person’s sentence proceedings. Finally, it was unclear whether the judge felt himself entitled to have regard to views formed during Ms Tarrant’s evidence in the course of Mr Roff’s trial when sentencing Ms Tarrant.

(i) sentencing hearing

1. [67]

   Finally, there was the sentencing hearing. Although the judge intended it to be a “joint sentencing hearing”, he accepted that the evidence of her trial would be evidence in her sentencing proceeding, but made no similar order with respect to her evidence given in Mr Roff’s trial. The psychiatrists were then recalled. The prosecutor called Dr Yvonne Skinner, who had prepared a supplementary report and gave further oral evidence. Of her evidence-in-chief, approximately one-half was led by the judge. On one view, the questions could be described as seeking clarification of the basis of the opinions expressed by Dr Skinner; on another view, the questioning was argumentative and challenged those conclusions. The questioning reached a stage where the judge asked: [56]

2. [68]

   The judge overruled Ms Manuell’s objection [57] and proceeded to question Dr Skinner about “love letters” sent by Ms Tarrant to Mr Roff. The implication of the questioning was that if the diagnosis of depression were to be accepted, it was unlikely that Ms Tarrant was truly in love with Mr Roff and looking forward to a new life with him; rather, she was dissembling and manipulating him.

(j) conclusions

1. [69]

   What precisely the lay observer might be thought to understand at this point depends upon the degree of sophistication and legal experience with which he or she is endowed. The observer should be taken to understand that (a) the criminal trial is an adversary process in which prosecutor presents the case for the State, to which the accused responds; (b) matters of fact and evaluation are for the jury; (c) the judge’s role at a trial is to remain impartial and above the fray; (d) no step should be taken with respect to an accused person or an offender which prejudices their interests in circumstances where (having a lawyer) they are unrepresented and not on trial; (e) the accused or offender is not expected to meet a case not presented by the prosecutor and (f) in sentencing, the judge must find facts, but not so as to undermine the jury’s verdict unfavourably to the offender.

2. [70]

   In our view the observer would have been troubled by (a) the judge’s attempt to establish levels of culpability depending upon the degree of impairment, (b) the questioning on a sentencing hearing of the underlying basis of the expert opinions which went to the jury unchallenged, and (c) a reaffirmation through the questioning of the opinions that the judge had earlier been expressing, albeit tentatively and argumentatively, which had not been put forward by the prosecution.

3. [71]

   The stage had been reached where the lay observer would give significant weight to the fact that the judge had taken over the leading of evidence from a prosecution witness. In the course of his questioning counsel for Ms Tarrant had objected to what appeared to be a challenge to the basis of the opinion that went to the jury. The judge overruled the objection, stating, “even if I should accept it was likely … that she was suffering from depression I would be entitled to have that evidence from this witness to ascertain how that marries with the other evidence in the case.” [58] The observer would know that the other evidence in the case was primarily the applicant’s evidence, which was consistent with that already supplied to the doctors when they prepared their reports.

4. [72]

   By that stage of the proceedings, the hypothetical lay observer might well have thought that the judge would not make findings of fact on the basis of the evidence given at trial, or the evidence presented by the prosecution, but rather on the basis of his own assessment of the psychiatric material. Where unchallenged evidence has been put before the jury by the parties, based largely on common ground and from experts whose qualifications were unchallenged, the lay observer might well think that the judge might have stepped beyond the role of the impartial arbiter of the facts as presented by the parties. That in our view would be sufficient to constitute a reasonable apprehension of bias. For that reason, the sentencing process has miscarried and must be repeated. The sentence imposed on Ms Tarrant must be quashed.

Sentencing judgment – other challenges

1. [73]

   In the event that the applicant were unsuccessful on her first ground, there were a number of further challenges to the sentencing judgment. Given that she has been successful on ground 1, it is not necessary to determine the further grounds. However, as the matter will be remitted to a judge who will not have had the benefit of having presided over Ms Tarrant’s trial and heard her evidence, it is desirable that this Court address aspects of the approach which underlay the first sentencing judgment.

2. [74]

   There is no doubt that a central feature of the sentencing exercise was for the judge to determine the culpability of the offender and place it within the broad spectrum covered by the offence of manslaughter. In undertaking that exercise, he was required to act consistently with the jury’s verdict. The verdict of manslaughter reflected a view formed by the jury with respect to the operation of s 23A of the Crimes Act.

3. [75]

   This was no easy task. The first stage of the exercise involved a series of steps which are best identified by reversing the structure of s 23A(1). Thus the jury had to be satisfied (on the balance of probabilities) that there was (a) an “underlying condition”, other than one of a “transitory kind,” (b) from which arose an “abnormality of mind”, (c) which led to “substantial impairment” of the person’s mental capacity identified in three respects. The three elements were (i) capacity to understand events, (ii) capacity to judge whether one’s actions are right or wrong and (iii) capacity for self-control.

4. [76]

   The jury had expert evidence from Dr Yvonne Skinner (instructed by the prosecutor) and Dr Carolyn Quadrio (instructed on behalf of the offender). It was common ground that their reports supported a finding of “substantial impairment” for the purposes of s 23A(1). They were not cross-examined; nevertheless, in his summing up, the trial judge emphasised that, having heard evidence going to the primary facts (which appeared as history in the expert reports) it was a matter for the jury to make their minds up as to whether the facts were as the psychiatrists opined, and whether the facts the jury found to have been established supported the psychiatric opinions. The jury, by their verdict, demonstrated that they must have substantially accepted the psychiatric opinions, because there was no other evidence upon which they could have reduced the conviction to manslaughter. Further, although in the sentencing judgment there was separate analysis of whether the impairment suffered by the applicant related to her capacity to understand events, or to judge right from wrong, or to control herself, the matter was left to the jury on the basis that “both psychiatrists must have intended to say that her capacity was substantially impaired in each of those respects.” [59]

5. [77]

   In short, there was no basis in the expert opinion evidence to challenge the elements of capacity addressed in the section, and it is far from clear that a lay person could properly draw alternative conclusions as to those matters without rejecting the findings of the psychiatrists. At least implicitly, the findings of the experts must have been accepted by the jury. Indeed, most of the factual circumstances which the judge relied on in sentencing as limiting the scope of the applicant’s impairment were put to the jury in the summing up as matters which the jury could and should assess in reaching their own view as to whether the impairment was substantial. The verdict implied general acceptance of the evidence.

6. [78]

   There is no doubt that a general verdict can be opaque with respect to the reasons for the ultimate conclusion. Usually, a verdict which results in a sentence being imposed is one which could properly be described as adverse to the offender. In this case, the jury must have been satisfied on the balance of probabilities that the applicant had made out her affirmative defence. Where that defence depends to a high degree on expert opinion evidence, the judge undertaking the sentencing exercise must have good reasons for supposing that there were elements of the evidence which were not accepted by the jury before undertaking his or her own analysis. Further, those reasons should be explained in the judgment. To adopt a different approach would be to risk not giving full weight to the fact that the jury had reached its verdict on the basis of the evidence and, to the extent that aspects of the evidence needed to be resolved, must have resolved them, at least in large part, favourably to the offender.

7. [79]

   Furthermore, the jury’s verdict was not merely an acceptance of the fact that the applicant’s mental capacities were substantially impaired by an abnormality of mind arising from the diagnosed pre-existing conditions (which included a persistent depressive disorder and a personality disorder with dependent traits); rather, they found that the impairment was “so substantial” as to warrant reduction of the offending from murder to manslaughter. That, in accordance with the section and with the jury directions, was a task which had been undertaken on the basis of “community values” as to levels of culpability, a test which cannot be reduced to component parts or specific standards. Yet, in carrying out the sentencing exercise, the judge said that the verdict merely set a threshold and that “[b]eyond that threshold, how severely her faculties were compromised must be determined by me in order to gauge the corresponding degree of reduction in culpability and the consequent impact on her sentence.” [60]

8. [80]

   The evidence and submissions on sentence were cast in terms of whether the degree of her substantial impairment was severe, moderate, or mild. It appears that evidence as to this kind of “range” was imposed on the psychiatric evidence as a result of developments in the course of the pre-sentence hearings. But if this exercise is to be worthwhile, it cannot be based on the degree of impairment above “substantial impairment” but must be based upon the highly imprecise finding that the impairment was “so substantial as to warrant liability for murder being reduced to manslaughter.” The very imprecision which must attach to the community values inherent in this judgment is inimical to any exercise in further gradations of the degree of impairment.

9. [81]

   In support of the passage as to the function of the sentencing judge, reference was made to R v Low [61] and R v Antaky. [62] As explained in Antaky, it is true that some forms of impairment may be “gross” and “some may only just fall within the description of ‘substantial’ so as to warrant the reduction.” [63] Nevertheless, it is to be doubted that these authorities support the approach adopted in the present case assuming fixed categories of degree. Furthermore, to describe the evidence as barely more than the minimum necessary to justify reduction of the verdict to manslaughter is to assume that the judge is in a position to evaluate the standard which the jury applied.

10. [82]

    Importantly in the present case, the judge found: [64]

11. [83]

    The judge stated that he accepted Dr Quadrio’s assessment that the relationship between the applicant and Rez “was a course of ‘physical and psychological and verbal abuse, threats of violence and both economic and social control’ in which the ‘critical element [was] the exercise of power and control’.” [65] He continued:

12. [84]

    The judge qualified his acceptance of this evidence: [66]

13. [85]

    To extract a phrase, seen to be of some importance, from a lengthy report, on which the author was not cross-examined, and dismiss her conclusion as not established, is fraught with difficulty. The judge’s summary did not do justice to the full description given by Dr Quadrio of the applicant’s mental condition and capacities and the reasons for her assessment. Secondly, despite finding that the history accepted by the psychiatrists was supported by the evidence of the applicant at trial, the judge set out to identify 12 instances of “violence” which occurred during their nine year relationship. He stated: “I accept her evidence as to these and find them proved on the balance of probabilities”. [67] Amongst the list of events, some involved multiple incidents. One was set out in the following terms: [68]

14. [86]

    Of the series of incidents, the judge said: [69]

15. [87]

    Much of this assessment was challenged on the appeal as factually inaccurate. In particular, it was noted that the last sentence was inconsistent with the history recorded by Dr Skinner that “physical abuse and emotional abuse became a lot worse after they moved to Dubbo in June 2012.” [70]

16. [88]

    Further, the judge said: [71]

17. [89]

    What then occurred is troubling. Relying primarily on evidence given at the sentencing hearing, the judge concluded that Dr Skinner (a) did not consider that the applicant’s capacity to understand events was impaired at all, [72] (b) gave evidence which “does not describe an impaired capacity to make a moral judgment”, [73] and (c) had found that the impairment of self-control “must be rated as minimal, amounting to a lack of self restraint with respect to pursuing her own objectives at the expense of another’s life.” [74]

18. [90]

    With respect to Dr Quadrio’s evidence, the judge stated:

19. [91]

    These findings cannot be reconciled with the findings which must have been made by the jury, albeit the present findings were largely based upon evidence given by the psychiatrists, to a significant extent at the insistence of the sentencing judge, for the purposes of sentencing. No attempt was made to reconcile these findings with the evidence before the jury; nor was it put to the psychiatrists on the sentencing hearing that the evidence they were then giving was inconsistent with their principal reports which had been left to the jury.

20. [92]

    There is, on occasion, some leniency accorded in a sentencing hearing to allow evidence given on behalf of the offender which is not consistent with that given at the trial. However, to rely upon evidence on sentence which is fundamentally inconsistent with evidence given at the trial (or at least to make findings based on it which are inconsistent with those which, in all probability, formed the basis of the jury’s verdict), in a manner favourable to the prosecution, should not be permitted.

21. [93]

    It is not necessary to say more in relation to the manner in which the sentencing proceeded and was determined. There will have to be a further hearing at which another judge will have to determine what evidence should be admitted beyond the evidence which went before the jury at the trial. It follows that on re-sentencing, the judge should not consider himself or herself bound to adopt any particular procedural steps or to pay regard to findings of fact which may have been made in the earlier sentencing proceeding.

Orders

1. [94]

   The Court makes the following orders:

   1. (1)

      Grant the applicant leave to appeal from the sentence imposed in the Common Law Division on 19 August 2016.

   2. (2)

      Allow the appeal and set aside the sentence and orders made by Fagan J in the Common Law Division.

   3. (3)

      Remit the matter to the Common Law Division for the sentencing of the applicant with respect to the offence of manslaughter of Alois Rez.