[2020] SGCA 2

This judgment text has undergone conversion so that it is mobile and web-friendly. This may have created formatting or alignment issues. Please refer to the PDF copy for a print-friendly version.

In the COURT OF APPEAL of the republic of singapore

[2020] SGCA 2

Criminal Reference No 6 of 2018

Between

Public Prosecutor

… Applicant

And

GCK

… Respondent

Criminal Motion No 7 of 2019

Between

Public Prosecutor

… Applicant

And

GCK

… Respondent

judgment

[Criminal Procedure and Sentencing] — [Criminal references] — [Court’s substantive jurisdiction to answer questions] — [Court’s power to reframe questions]

[Evidence] — [Proof of evidence] — [Standard of proof] — [Beyond a reasonable doubt]

[Evidence] — [Proof of evidence] — [Standard of proof] — [“Unusually convincing” standard]

[Evidence] — [Witnesses] — [Eyewitness testimony]

This judgment is subject to final editorial corrections approved by the court and/or redaction pursuant to the publisher’s duty in compliance with the law, for publication in LawNet and/or the Singapore Law Reports.

Public Prosecutor

GCK and another matter

[2020] SGCA 2

Court of Appeal — Criminal Reference No 6 of 2018 and Criminal Motion No 7 of 2019
Sundaresh Menon CJ, Andrew Phang Boon Leong JA and Judith Prakash JA
25 September 2019

22 January 2020 Judgment reserved.

Sundaresh Menon CJ (delivering the judgment of the court):

Introduction

1 There is but one true evidentiary standard in criminal law – proof beyond a reasonable doubt. This sacrosanct principle is easy to express, but defies simple definition. Difficulties often arise where the sole basis for a conviction is the uncorroborated testimony of a single witness. These difficulties may intensify where, as in the present case, that witness is not the victim (who was mentally unfit to testify), but an eyewitness. The case law has developed techniques to manage these difficulties. One, in particular, is the stipulation that the uncorroborated evidence of a victim must be “unusually convincing” if it is to be accepted as the sole basis for convicting an accused person. But just what does that mean? And should it be applied to the evidence of an eyewitness, and if so, how? In our judgment, the difficulties are considerable, but they are not insurmountable. What is required in the final analysis is a careful and holistic consideration of the evidence. At all times, the court remains concerned with assessing whether the Prosecution has met its burden of proving its case beyond a reasonable doubt, no more and no less.

2 The present case affords us the opportunity to clarify certain important aspects of the law of evidence in the foregoing context. The respondent (“the Respondent”), who was a male employee at a nursing home (“the Home”), was charged with one count of outrage of modesty (“the OM Charge”) under s 354(1) of the Penal Code (Cap 224, 2008 Rev Ed) (“the Penal Code”). The alleged victim (“the Victim”) was an elderly female resident at the Home who was suffering from serious physical and cognitive disabilities. She was found unfit to testify. The Prosecution’s case therefore rested substantially upon the testimony of an eyewitness – Nurse MJ, a female nurse at the Home – who alleged that she had seen the Respondent straddling the Victim with his trousers pulled down and the Victim’s diaper exposed, and with his groin placed on the Victim’s groin.

3 In the District Court, the Respondent was convicted of the OM Charge. On appeal, the High Court reversed the decision of the District Court and acquitted the Respondent. In so doing, the High Court appeared to draw a distinction between the evidence of alleged victims on the one hand, and eyewitness testimony on the other. This led the Prosecution to file Criminal Reference No 6 of 2018 (“the present Criminal Reference”), raising before us several questions specifically pertaining to the “unusually convincing” standard and its application to sexual and non-sexual offences as well as to the evidence of alleged victims and eyewitnesses. As shall be seen, the remit of those questions also touched more generally upon the meaning of the standard of proof beyond a reasonable doubt.

Background facts

4 The Respondent was 32 years old at the material time. He had been working at the Home since 2010, initially as a housekeeping attendant. After suffering a back injury in 2013, the Respondent stopped working in that capacity, and instead undertook maintenance and cleaning duties as a maintenance technician for the Home. Those were his duties at the time of the alleged incident on 26 November 2016.

5 The Victim was 55 years old at the material time. She had suffered a series of strokes some years earlier, which limited her mobility on the left side of her body and impeded her speech. As a result of those strokes, the Victim displayed easily-changing moods, alternating between crying and giggling. She was generally unable to raise her voice, and would make high-pitched cries of a soft to moderate volume. Due to her cognitive disabilities, the Victim was certified unfit to testify at the trial.

6 The Victim occupied Bed 7 of a room that was exclusively for female residents (“the Room”). The Room was located on one of the upper levels of the Home and contained 12 beds. It was partially divided by a wall into an inner section and an outer section. The Victim was bed-bound, and her bed was located at the far corner of the inner section of the Room next to the windows. For ease of reference, a repurposed sketch of the Room is set out at Annex A to this judgment.

7 The Prosecution’s case at the trial and in the appeal depended principally on the testimony of Nurse MJ, who had been working at the Home since April 2016. The Victim was one of the residents under Nurse MJ’s nursing care.

Accounts of the events on the day of the alleged offence

Mdm RM’s, Mr ST’s, and Dr S’s testimonies

8 On 26 November 2016, the Home was hosting a community involvement programme on the ground floor. The programme was scheduled to last from 2.00pm to 4.00pm. An administrative officer of the Home, Mdm RM, testified that all of the Home’s staff, including its maintenance staff, would have been made aware of the programme through the noticeboards. Mdm RM further testified that the Respondent was the only maintenance staffer on duty that day. At around 1.00pm, the Respondent assisted Mdm RM in setting up audio-visual equipment for the programme.

9 Mr ST, the director of the Home, testified that five residents in the Room (including the Victim) remained in their beds during the community involvement programme. They did not join in the programme because they were asocial, required assistance in movement, or were suffering from mental disabilities that rendered them unable to mingle with members of the public.

10 Dr S, who was the resident physician of the Home, opined that the other four residents who were present in the Room at the material time had dementia, amnesia, or were otherwise incapable of communicating a narrative. In his view, none of them were mentally capable of testifying.

Nurse MJ’s testimony

11 Nurse MJ testified that she started her shift at 7.00am on 26 November 2016. At around 3.41pm, she proceeded on her rounds to the Room.

12 Nurse MJ noticed that the curtains around the beds in the inner section of the Room were fully drawn, except for the curtains around the Victim’s bed (Bed 7), which were half-drawn. She found this odd as the curtains were usually only drawn when the residents’ diapers were being changed. However, none of the residents in the Room was having her diaper changed at that time. In fact, several of the residents were not even occupying their beds, as they were either at the dining hall or at the community involvement programme. Aside from the Victim, the only other resident present in the inner section of the Room was Mdm MG in Bed 6, located across from the Victim.

13 Nurse MJ walked towards Mdm MG to check on her. As she was drawing open the curtains to Bed 6, she heard a crying sound emanating from Bed 7. Nurse MJ recognised this sound as one that the Victim would make whenever she was being moved or was in pain.

14 When Nurse MJ turned, she saw that the curtains to Bed 7 were half-closed. She testified that she had a full view of Bed 7 because the curtains were not drawn across the bed’s width and she was only about one-and-a-half arm’s lengths away. She saw the Respondent on the bed with his knees astride the Victim. The Respondent’s pants were lowered and Nurse MJ saw his exposed buttocks. The Victim’s pants were also lowered and the left side of her diaper was open.

15 Nurse MJ was shocked. She testified that “[the Respondent’s] groin area and [the Victim’s] groin area w[ere] together”, and she thought “it was something to do with sex and it was wrong”. She had a half-view of the Respondent’s face and recognised him as an employee of the Home. The Victim continued to make the crying sound. Nurse MJ observed the scene for about five seconds before leaving. She did not attempt to stop the Respondent because she was frightened.

16 Nurse MJ was not aware if the Respondent had noticed her. She did not know why the Respondent had entered the Room. It was the Home’s policy that male staffers had to be escorted by a female staffer when entering a room occupied exclusively by female residents.

17 The timing of the events recounted by Nurse MJ was consistent with what was recorded on the close-circuit television (“CCTV”) situated at the entrance to the Room. The CCTV footage showed that Nurse MJ entered the Room at 3.41.32pm and remained there for a total of 11 seconds.

Nurse DS’s testimony

18 Nurse MJ immediately headed to the dining hall on the same level of the Home and called out to a male nursing aide, Nurse DS, who subsequently testified that Nurse MJ’s tone and appearance were not normal and that she sounded as if there was an emergency. On the first two times Nurse MJ called to him, Nurse DS told her to wait as he was tending to a patient, but on the third time, she shouted at him and asked him to “[p]lease go and see what [the Respondent] is doing on [the Victim’s] bed”.

19 The CCTV footage showed that about a minute and 40 seconds after Nurse MJ left the Room, Nurse DS entered. Nurse DS saw that the curtains around Bed 8 were fully drawn, which he too found strange as that would only be done when the residents’ diapers were being changed. Standing on tiptoe, Nurse DS was able to peer through the upper netting portion of the curtains around Bed 8 to get a view of Bed 7. He saw the Respondent kneeling on the floor in the space between Bed 7 and Bed 8 and apparently looking at his mobile phone. The curtains around Bed 7 were fully open, and Nurse DS observed that the Victim was sleeping on the bed. She appeared normal and there were no sounds coming from her.

The Respondent’s testimony

20 The Respondent claimed that sometime around lunchtime on 26 November 2016, one of the Room’s residents, Mdm JP, asked him to repair her portable television as he had done on previous occasions. At around 3.30pm, the Respondent remembered Mdm JP’s request and decided to attend to it in the Room. He claimed not to have known of the Home’s policy forbidding male staffers from entering any room occupied exclusively by female residents without being accompanied by a female staffer.

21 The Respondent headed to Mdm JP’s bed, Bed 8, which was adjacent to the Victim’s bed, Bed 7. Mdm JP was not present at that time. The Respondent knelt on the floor between Bed 7 and Bed 8, and changed the fuse for Mdm JP’s portable television. He switched on the power for the television and was adjusting the channels when he heard a sound from Bed 7, which sounded as though “someone was tapping on the [railing] of the bed”.

22 The Respondent claimed to have turned and seen the Victim’s head touching the side railing of her bed. He noticed that her pillow was displaced. He observed tears flowing from the Victim’s eyes and thought the Victim must be in pain as her head was bent towards the railing. There was another round pillow to the side of the Victim. The Respondent claimed to have placed his left knee between the bars of the side railing to reach for the round pillow. No part of his body touched the Victim while he was reaching for the round pillow. He placed the round pillow under the Victim’s head. The Victim then smiled a little.

23 As the Respondent returned to adjusting the channels on Mdm JP’s portable television, he noticed some “food greens” consisting of rice grains and crushed biscuits on Mdm JP’s bed. He proceeded to the toilet to dispose of those items. He then returned to the Room and placed Mdm JP’s television on her bedside table. At about the time he replaced the television, a friend of Mdm JP, Ms SBR, came by Bed 8 to retrieve Mdm JP’s spectacles for her. The Respondent then left the Room. He had not noticed Nurse MJ or Nurse DS entering the Room.

Events subsequent to the alleged incident

24 Nurse MJ left work at the end of her shift. She then telephoned a senior staff nurse, SSN JS, because she remained troubled by what she had seen despite Nurse DS having told her that he had seen nothing amiss. SSN JS met Nurse MJ to discuss the incident, and SSN JS then called Mr ST. The Victim was not sent for an immediate medical examination. Instead, Mr ST interviewed the Victim on the following day (27 November 2016), and viewed the CCTV footage on 28 November 2016. The matter was then escalated to management.

The police investigations

25 On 23 January 2017, a police report was lodged against the Respondent, who was arrested on the same day.

26 The Respondent was interviewed on 26 January 2017 by Assistant Superintendent Razali bin Razak (“ASP Razali”) under s 22 of the Criminal Procedure Code (Cap 68, 2012 Rev Ed) (“the CPC”). In contrast to his oral testimony at the trial that he had placed only his left knee on Bed 7 while he was reaching for the Victim’s round pillow, the Respondent claimed in his statement to ASP Razali on 26 January 2017 (the “First Statement”) that “[a]s the bed was too high and difficult for [him] to retrieve the … pillow”, he had “placed both [his] knees at the left side of the bed and [taken] the pillow” [emphasis added]. A drawing of Bed 7 was also appended to the First Statement, with two line markings made on the Victim’s left side to indicate where the Respondent had allegedly placed his knees.

27 ASP Razali interviewed the Respondent again on 30 January 2017 and recorded another statement from him (the “Second Statement”). Among other things, the Respondent said as follows in the Second Statement:

The two “/ /” markings was [sic] where I placed my knee on the side of the bed. … [emphasis added]

28 In his evidence-in-chief, the Respondent claimed that in the Second Statement, he was referring to only one of his knees. He explained that after he had given his First Statement, he had spoken to some of the other remandees in the lock-up. They “gave [him] courage and told [him] that [he should] just tell what actually happened, nothing to fear. Then [he] started thinking and [he] could recollect what actually happened.”

29 In contrast, ASP Razali testified that the reference to “my knee” in the Second Statement was a typographical error, and that the phrase ought to refer to both the Respondent’s knees.

30 On 31 January 2017, the day after the Respondent’s Second Statement was recorded, ASP Razali escorted the Respondent to the Home for a scene investigation (“the Scene Investigation”). A photograph, Exhibit P19, was taken on ASP Razali’s directions. P19 showed two arrows placed on Bed 7. At the trial, the Respondent agreed that the arrows corresponded to both his knees, and that he had indicated as such to ASP Razali during the Scene Investigation. He claimed that he had mistakenly told ASP Razali that he had placed both his knees on the bed because he was “scared” at that time, having been told by some of the remandees in the lock-up that he was facing “a very dangerous allegation”. He had not informed ASP Razali of this mistake because he was “afraid that if [he] tell that [he] had made a mistake, the officer would be angry with [him]”.

The trial in the District Court

31 The learned trial judge in the District Court (“the District Judge”) convicted the Respondent of the OM Charge: see Public Prosecutor v GCK [2018] SGDC 195 (“the DJ’s GD”).

32 Given that the Victim was unfit to testify and that the offence which the Respondent was charged with was a sexual offence, the District Judge applied the “unusually convincing” standard affirmed by this court in Public Prosecutor v Mohammed Liton Mohammed Syeed Mallik [2008] 1 SLR(R) 601 (“Mohammed Liton”) at [39] to scrutinise the evidence of Nurse MJ, who was an independent witness to the alleged offence: see the DJ’s GD at [50]. In this connection, the District Judge found that Nurse MJ’s evidence was internally consistent. She had not shouted or reacted at the time because she was afraid. Her demeanour showed that she was deeply affected by what she had seen. The District Judge rejected the Respondent’s claim that Nurse MJ had fabricated the allegations against him. Nurse MJ’s evidence was consistent with and supported by the CCTV footage and Nurse DS’s testimony. The District Judge accepted that the Respondent was not sexually assaulting the Victim when Nurse DS walked into the Room, but concluded that he had stopped doing so after sensing Nurse MJ’s presence a few moments earlier. The District Judge also concluded that the Respondent had left the curtains to Bed 7 partially open in order to sense if someone was approaching because he did not want to be caught red-handed should someone suddenly draw open the fully-closed curtains: see the DJ’s GD at [53]–[59].

33 As against this, the District Judge found the Respondent’s evidence internally inconsistent. The Respondent could not provide a consistent explanation for why he had said in his First Statement that he had placed both his knees on Bed 7 while he was reaching for the Victim’s pillow, when he later claimed that he had placed only one knee; nor could he say what the “mistake” in his First Statement was, and what he was scared of so as to have made such a “mistake”. The District Judge concluded that the Respondent had belatedly realised that his account of having placed both his knees on Bed 7 was “highly unnatural and contrived”, and had therefore sought to change his story and claim that he had placed only one knee on the bed: see the DJ’s GD at [62]–[63].

34 During the trial, the District Judge conducted a scene visit to the Home (“the Trial Scene Visit”), where he requested the Respondent to demonstrate how he had allegedly retrieved the Victim’s round pillow with one knee placed on Bed 7. The District Judge thought this posture too was “highly unnatural and contrived”. He then requested the Respondent to reach for the pillow without placing either of his knees on the bed, and found that given the Respondent’s height, he could easily reach for the pillow without placing either knee on the bed. Photographs of the demonstration taken during the Trial Scene Visit were adduced as Exhibits P91 and P94 at the trial: see the DJ’s GD at [63].

35 Mdm JP was called by the Defence to corroborate the Respondent’s account that she had asked him to repair her portable television on 26 November 2016. However, the District Judge rejected her evidence. He noted that in her statement to the police, Mdm JP had said that the only time she had requested the Respondent to repair her television was in June or July 2016, and that she had not made such a request on 26 November 2016. She was not able to satisfactorily explain why she had given such a contradictory account in her statement to the police. In any case, the District Judge noted that even if the Respondent had indeed gone to the Room to attend to Mdm JP’s request, this did not have any bearing on whether he had also committed the offence alleged against him: see the DJ’s GD at [69]–[72] and [75].

36 After taking into account the seven weeks the Respondent had spent in remand, the District Judge meted out a sentence of 22 months’ imprisonment for the OM Charge, which was close to the maximum two years’ imprisonment stipulated in s 354(1) of the Penal Code. He also imposed three strokes of the cane: see the DJ’s GD at [2] and [98]–[99].

The magistrate’s appeal in the High Court

37 The Respondent appealed against both his conviction and his sentence by way of Magistrate’s Appeal No 9156 of 2018 (“MA 9156”), which was heard by the learned High Court judge (“the Judge”). On 23 November 2018, the Judge issued his oral grounds acquitting the Respondent of the OM Charge (the “Oral Grounds”).

38 In his Oral Grounds, the Judge observed as follows:

3 … [T]he court is … to require that in the absence of objective corroborative evidence, … the evidence of the complainant against the alleged attacker is unusually convincing.

4 In the present case, while I see no reason to doubt that the witness, [Nurse] MJ, was honest, I could not find that her evidence alone was sufficient to convict the [Respondent]:

(a) The evidence of the assault was not so clear and definitive that it carried strength and cogency in its description. Her viewing of the incident was, perhaps not fleeting, but neither was it sufficient to be definitive.

(b) Her reaction and subsequent behaviour could be both the result of her seeing an actual assault, or her being mistaken about seeing such an assault[.]

…

5 The reporting to others could not assist the prosecution’s case ultimately; it is weaker than reporting by a victim, who would be recounting an assault on herself. A victim is not likely to have misperceived what was happening to her. But in contrast, a witness reporting what she saw would still be subject to mistake or misapprehension.

39 On 30 November 2018, the Prosecution filed the present Criminal Reference. On 27 February 2019, the Judge issued his written grounds: see GCK v Public Prosecutor [2019] SGHC 46 (the “Written GD”). In the Written GD, the Judge elaborated on his decision to acquit the Respondent of the OM Charge in the following crucial passages:

33 A third-party witness who is disinterested and neutral may not appear to require the cautionary approach embodied by the need for evidence to be unusually convincing. Nonetheless, as the evidence remains word against word, a similar imperative applies: the court should be slow to convict in the absence of unusually convincing evidence. Hence, the District Judge was not wrong in transposing the unusually convincing requirement to an eye-witness. But even if that requirement were not applied to Nurse MJ, her evidence would not have been sufficient to secure a conviction simply because it fell short of the required cogency and strength that on any non-sexual case would be required for a conviction to stand: a reasonable doubt that she was mistaken could not be excluded. So on either the application of the unusually convincing requirement, or otherwise, the evidence for the prosecution did not make out a case beyond a reasonable doubt.

…

35 While there is no reason to doubt that Nurse MJ was truthful and candid, her evidence of the [Respondent’s] sexual assault on the [V]ictim was not definitive and conclusive of the matter, and there were gaps in the evidence that were not adequately addressed. …

36 Nurse MJ was a bystander. She was not the victim. While an eye witness’s testimony has the advantage of being an independent account of the events, eye witness evidence is always subject to possible misapprehension and errors in observation. In contrast to a victim’s own testimony, an eye witness’s account would be subject to a greater degree of misperception, misapprehension and misattribution. This is particularly so where the incident might have taken place over a period of time and the eye witness was only present for a brief moment. Without an appreciation of the full context in which the events unfolded, a bystander’s account may be liable to misinterpretation. An eye witness’s testimony is not a recording. …

…

39 Nurse MJ had only about a five-second glimpse of the alleged assault. She may have indeed seen something, or she may have been mistaken. The possibility of mistake or misapprehension is higher the shorter the observation.

40 The Judge also noted that Nurse DS had entered the Room shortly after Nurse MJ left, and had observed the Victim to be asleep and looking normal. The “drastic change” in the Victim from crying in pain to being asleep could not be lightly regarded, and there was also a question of the amount of time it would have taken for the Respondent to adjust his and the Victim’s clothes after the alleged assault. The Judge was further of the view that the District Judge had erred in discrediting Mdm JP’s testimony: see the Written GD at [40]–[41] and [47].

41 The Judge therefore acquitted the Respondent of the OM Charge.

The Prosecution’s motion to reframe the question posed

42 Before we turn to the present Criminal Reference, we address a preliminary issue pertaining to the question posed. The Prosecution had initially framed the question in the present Criminal Reference in these terms (“the Original Question”):

Question

In evaluating the uncorroborated evidence of a complainant in a sexual offence, is the evidence of an eye-witness to the alleged crime to be assessed by a higher standard than the “unusually convincing” standard which is applied to assess the evidence of the victim?

[emphasis added]

43 On 24 May 2019, the Prosecution filed Criminal Motion No 7 of 2019 (“CM 7”) to reframe the Original Question as follows (“the Reframed Question”):

What is the test to be applied when evaluating the evidence of an independent and honest eye-witness to a sex crime (“the independent eye-witness”), where such evidence – (a) is uncorroborated; and (b) forms the sole basis for a conviction? In particular:

(1) Is the evidence of the independent eye-witness inherently less reliable than that of the alleged victim?

(2) Is the test to be applied the “unusually convincing” test as laid down in Khoo Kwoon Hain v PP [1995] 2 SLR(R) 591, or the test for identification evidence as laid down in Heng Aik Ren Thomas v PP [1998] 3 SLR(R) 142, or some other test?

44 The Prosecution explained that the Original Question had been formulated in light of the Judge’s Oral Grounds. However, after having sight of the Judge’s Written GD, it considered that the Reframed Question would better reflect and clarify the relevant issues of law of public interest that it believed had arisen. In the circumstances, the Prosecution sought the exercise of this court’s powers under s 397(4) of the CPC to reframe the Original Question as the Reframed Question. Section 397(4) of the CPC states:

In granting leave to refer any question of law of public interest under subsection (1), or where the Public Prosecutor refers any question of law of public interest under subsection (2), the Court of Appeal may reframe the question or questions to reflect the relevant issue of law of public interest, and may make such orders as the Court of Appeal may see fit for the arrest, custody, or release on bail of any party in the case.

45 In his written submissions, Mr Lau Wen Jin (“Mr Lau”), counsel for the Respondent, took the position that CM 7 should be dismissed. However, at the oral hearing before us, Mr Lau indicated that he would not object to the reframing of the Original Question, though he nevertheless reserved his position as to whether this court should answer the Reframed Question in any event.

Our decision on the Prosecution’s application to reframe the Original Question

46 In Bachoo Mohan Singh v Public Prosecutor and other applications [2010] 1 SLR 966 (“BMS (No 3)”) at [38], V K Rajah JA indicated that “the Court of Appeal, in deciding the [question] of law of public interest reserved by the High Court, has the power to reframe the question to achieve clarity”. In this regard, Rajah JA cited the Court of Appeal’s observations in Public Prosecutor v Fernandez Joseph Ferdinent [2007] 4 SLR(R) 1 (“Fernandez”) at [19]:

… [A] refashioning of a question being posed by an applicant to this court in a criminal reference is neither novel nor inappropriate. The overriding task of this court in any criminal reference is to clarify questions of law of public interest. … [W]here a question is couched in a manner which would inadvertently mask its true import (which is the situation here), the court retains a discretion to pose the question in a manner which will be more appropriate and which will ensure the substance of the question is rendered clear, save that the refashioned question has to remain within the four corners of s 60 of the [Supreme Court of Judicature Act (Cap 322, 1999 Rev Ed)] …

47 The relevant observations in BMS (No 3) and Fernandez were made in the context of, respectively, s 60 of the Supreme Court of Judicature Act (Cap 322, 2007 Rev Ed) and s 60 of the Supreme Court of Judicature Act (Cap 322, 1999 Rev Ed), which were the predecessor statutory provisions of s 397 of the CPC. We are satisfied that they are equally applicable to s 397 of the CPC. The overall objective of a criminal reference is to clarify questions of law of public interest, and this would be frustrated if the questions are unclearly articulated. In our judgment, questions which are too narrowly framed might obscure the pertinent public interest elements. Conversely, questions which are too ambiguous or broad would prevent the parties from focusing their minds on addressing the relevant issues should the criminal reference be heard: see BMS (No 3) at [86].

48 This is not to say that this court should entertain requests to reframe the question posed in every case. When it is clear that no question of law of public interest arises, then reframing the question originally posed would be a fruitless exercise. Nor does it mean that having reframed a question, this court is bound to answer it in the exercise of its substantive jurisdiction. This court ultimately retains a broad discretion guided by the statutory requirements of s 397 of the CPC. In the present case, we accept the Prosecution’s explanation that the Original Question had been framed with reference to the Judge’s Oral Grounds, when it did not have the benefit of the Judge’s Written GD, and it was the latter that the Reframed Question was predicated upon.

The Applicable Standard Question

49 The Prosecution submitted that the Judge’s Oral Grounds at [3]–[4] did not make clear whether the “unusually convincing” standard was to be applied when the Prosecution was relying solely on the uncorroborated evidence of an eyewitness. This led the Prosecution to suggest in the Original Question that a higher standard than the “unusually convincing” standard was being applied by the Judge. In the Written GD at [33], the Judge then indicated that “even if [the ‘unusually convincing’ standard] were not applied to Nurse MJ, her evidence would not have been sufficient to secure a conviction simply because it fell short of the required cogency and strength that on any non-sexual case would be required for a conviction to stand” [emphasis added].

50 The learned Deputy Public Prosecutor, Mr Kow Keng Siong (“Mr Kow”), submitted that the Judge’s formulation gave rise to confusion as to whether one or more legal standards was to be applied when the uncorroborated evidence of an eyewitness formed the sole basis for a conviction. It was also unclear whether different legal standards would apply to sexual and non-sexual offences. This confusion gave rise to what we term “the Applicable Standard Question”.

51 In addition, Mr Kow submitted that the Judge had injected further uncertainty by alluding to the fact that the legal test(s) should apply even where an independent eyewitness was found to be “honest”, “truthful and candid”.

52 While we broadly agree with the Prosecution’s reformulation of the Original Question, we do not think the Applicable Standard Question should narrowly focus on purportedly “independent” eyewitnesses who are deemed to be honest witnesses, nor should it be confined simply to sexual offences. Mr Kow accepted that the assessment of a witness’s credibility is a separate inquiry, whereas it appears to us that the crux of the present inquiry is the reliability of an eyewitness’s observations. In addition, the underlying inquiry is also focused on the uncorroborated nature of the eyewitness’s evidence. As we shall go on to elaborate, this is not premised on whether the offence in question is of a sexual or a non-sexual nature.

53 In the circumstances, we reformulate the Applicable Standard Question as follows:

[Applicable Standard Question] What is the standard to be applied when evaluating the evidence of an eyewitness to a crime, where such evidence – (a) is uncorroborated; and (b) forms the sole basis for a conviction?

We note that the Prosecution’s proposed answer to the Applicable Standard Question is that the sole standard to be applied should be the “unusually convincing” standard.

The Inherent Reliability Question

54 Mr Kow turned to the portions of the Oral Grounds at [5] and the Written GD at [36] where the Judge appeared to suggest that the testimony of an eyewitness would be “weaker than reporting by a victim” and more prone to mistake or misapprehension. The Prosecution queried this aspect of the Judge’s reasoning, submitting that it might have the effect of raising the evidentiary threshold in cases involving the uncorroborated evidence of eyewitnesses. This led the Prosecution to pose the question whether “the evidence of the independent eye-witness [is] inherently less reliable than that of the alleged victim” (see [43] above).

55 Leaving aside the issue of the independence of eyewitnesses (see [52] above), we agree that the aforesaid portions of the Judge’s Oral Grounds and his Written GD raise what we term “the Inherent Reliability Question”, which we formulate as follows:

[Inherent Reliability Question] Is the evidence of an eyewitness inherently less reliable than that of an alleged victim?

The Specific Test Question

56 Mr Kow suggested that if we agreed that the answer to the Applicable Standard Question was that the “unusually convincing” standard should apply to the uncorroborated evidence of an eyewitness, it would be of public interest to consider how that standard should be applied. Mr Kow submitted that in applying that standard, the applicable test was the test for identification evidence established in Heng Aik Ren Thomas v Public Prosecutor [1998] 3 SLR(R) 142 (“Thomas Heng”), but also invited us to consider whether it might be “some other test”. In other words, we were asked to determine what the specific content of the “unusually convincing” standard should be in relation to the uncorroborated evidence of an eyewitness. We term the Prosecution’s query here “the Specific Test Question”.

57 In this regard, Mr Lau for the Respondent likewise accepted that the sole applicable standard for evaluating the uncorroborated evidence of an eyewitness was the “unusually convincing” standard. Without pre-empting the court’s decision on this issue in any way, we agree with the parties that the “unusually convincing” standard applies. But, for reasons we shall elaborate upon below, we are not persuaded that the applicable test is necessarily the Thomas Heng test. This unduly narrows the scope of the court’s evaluation of the evidence. We therefore elect to reframe the Specific Test Question as such:

[Specific Test Question] How should the court assess the evidence when it applies the “unusually convincing” standard?

58 Accordingly, we allow the Prosecution’s application in CM 7 to reframe the Original Question, though we also exercise our discretion to reformulate the Reframed Question in the manner elucidated at [53], [55], and [57] above. We turn to the next stage of the inquiry, which is whether we should exercise our substantive jurisdiction to answer that question.

The Court of Appeal’s substantive jurisdiction to answer a criminal reference

59 The present Criminal Reference was brought by the Prosecution under s 397 of the CPC, which reads:

Reference to Court of Appeal of criminal matter determined by High Court in exercise of its appellate or revisionary jurisdiction

397.—(1) When a criminal matter has been determined by the High Court in the exercise of its appellate or revisionary jurisdiction, and a party to the proceedings wishes to refer any question of law of public interest which has arisen in the matter and the determination of which by the Judge has affected the case, that party may apply to the Court of Appeal for leave to refer the question to the Court of Appeal.

(2) The Public Prosecutor may refer any question of law of public interest without the leave of the Court of Appeal.

…

(6) For the purposes of this section, each of the following is deemed to be a question of public interest:

(a) any question of law regarding which there is a conflict of judicial authority;

(b) any question of law that the Public Prosecutor refers.

60 As it was the Public Prosecutor who brought the present Criminal Reference, leave was not required under s 397(2) of the CPC. However, as we indicated in Public Prosecutor v Lam Leng Hung and others [2018] 1 SLR 659 (“Lam Leng Hung”) at [50], s 397(2) does not affect this court’s substantive jurisdiction to determine whether to answer the questions referred to it. In other words, while leave is not required where the questions are referred by the Public Prosecutor, the Court of Appeal is not invariably bound to answer the questions placed before it.

61 Next, we note that s 109(b) of the Criminal Justice Reform Act 2018 (Act 19 of 2018) recently amended s 397(6) of the CPC. In particular, in addition to questions involving a conflict of judicial authority (sees 397(6)(a)), questions of law that are referred by the Public Prosecutor are now also deemed to be questions of public interest by virtue of s 397(6)(b).

62 In our judgment, the effect of s 397(6)(b) is to expand the categories of questions of law that would be deemed to be questions of public interest. However, like s 397(2), this deeming provision does not impinge on this court’s substantive jurisdiction to decide whether to answer the questions referred to it. Where the exercise of this court’s substantive jurisdiction is concerned, we are guided by this court’s prior deliberations in Public Prosecutor v Goldring Timothy Nicholas and others [2014] 1 SLR 586. There, it was held at [26]:

… When exercising its substantive jurisdiction under s 397, the Court of Appeal will necessarily consider whether the case before it falls truly within the scope of that particular provision. This, in turn, entails considering whether all the requirements in s 397(1) are made out. … [emphasis in original]

63 Similarly, in Lam Leng Hung, this court stated at [50]:

… In the exercise of its substantive jurisdiction under s 397, the Court of Appeal must be satisfied that the application properly falls within the scope of the provision, and this in turn entails a consideration of whether the requirements in s 397(1) are made out (PP v Lim Yong Soon Bernard [2015] 3 SLR 717 at [16]; PP v Goldring Timothy Nicholas [2014] 1 SLR 586 at [26]). [emphasis in original]

64 In our judgment, ss 397(2) and 397(6)(b) serve to facilitate the bringing of a question before the Court of Appeal by the Public Prosecutor by, respectively, dispensing with the need to obtain leave and deeming any question of law referred by the Public Prosecutor to be a question of public interest. But the conditions which must be present for the court to answer the question referred to it remain unchanged. These four conditions were identified in Mohammad Faizal bin Sabtu and another v Public Prosecutor and another matter [2013] 2 SLR 141 (“Mohammad Faizal”) at [15] as follows:

(a) first, the reference to the Court of Appeal can only be made in relation to a criminal matter decided by the High Court in the exercise of its appellate or revisionary jurisdiction;

(b) second, the reference must relate to a question of law, and that question of law must be a question of law of public interest;

(d) fourth, the determination of the question of law by the High Court must have affected the outcome of the case.

Whether the conditions under s 397 were met

65 Mr Lau submitted that this court should not, in its substantive jurisdiction, answer the Reframed Question. In essence, Mr Lau’s objections were targeted at the third and fourth conditions set out in Mohammed Faizal, which he submitted had not been fulfilled in the present case.

66 In examining Mr Lau’s objections, it would be useful to set out in full the Reframed Question as we have reformulated it:

[Inherent Reliability Question] Is the evidence of an eyewitness inherently less reliable than that of an alleged victim?

[Specific Test Question] How should the court assess the evidence when it applies the “unusually convincing” standard?

Whether questions of law of public interest had arisen from the Judge’s decision

67 With regard to the Applicable Standard Question, Mr Lau submitted that no legal controversy had arisen because the Judge had applied the “unusually convincing” standard to assess Nurse MJ’s evidence. Similarly, Mr Lau contended that no controversy had arisen as to the Inherent Reliability Question because the Judge, in drawing a distinction between the evidence of an eyewitness and that of a victim, was simply making a “general starting proposition” which was “not an absolute statement, but a statement of relativity”.

68 As for the Specific Test Question, Mr Lau submitted that since the Judge had not departed from the “unusually convincing” standard in assessing Nurse MJ’s evidence, there was no need for the Judge to incorporate the factors enumerated in Thomas Heng into his analysis.

(1) Questions of law arising as to the Applicable Standard Question and the Specific Test Question

69 We are not persuaded that the Judge had simply and in a straightforward manner applied the “unusually convincing” standard to Nurse MJ’s evidence. It is true that the Judge had at [37] of the Written GD tentatively “agreed with the District Judge that the ‘unusually convincing’ standard may be applied to both victim and third-party eye witness accounts” [emphasis added]. However, several difficulties arise from his observations at [33] of his Written GD (see [39] above), which we reproduce again in full below:

70 We consider that these observations of the Judge give rise to significant ambiguity. For one, the Judge’s use of the phrases “even if” and “on either … or otherwise” suggests that he had in mind two different standards that could be applied to the evidence of an eyewitness or a victim, as the case may be. For another, the Judge’s allusion to “any non-sexual case” suggests that the “unusually convincing” standard is only confined to sexual offences. The foregoing gives rise to a question of law of public interest as to what the correct standard(s) should be when the uncorroborated evidence of an eyewitness forms the sole basis for a conviction.

71 We should stress that our holding is not meant to constrain judges from making obiter observations or findings in the alternative. On the contrary, judges should feel free to do so. However, where the central legal standard that is being applied in a case is unclear, and it injects ambiguity into an admittedly novel area of law (namely, whether the “unusually convincing” standard applies to an eyewitness’s uncorroborated testimony), a question of law of public interest would generally be found to have been raised.

72 We would further add that the distinction drawn by the Judge between sexual and “non-sexual case[s]” appears to sit uneasily with other authorities in the High Court to the effect that the “unusually convincing” standard is not confined to sexual offences (though admittedly in the context of the evidence of alleged victims rather than that of eyewitnesses): see Kwan Peng Hong v Public Prosecutor [2000] 2 SLR(R) 824 (“Kwan Peng Hong”) at [30], and XP v Public Prosecutor [2008] 4 SLR(R) 686 (“XP”) at [33]. The Judge’s holding thus created a conflict of judicial authority, giving rise to a further basis for finding a question of law of public interest: see s 397(6)(a) of the CPC; see also Lam Leng Hung at [58].

73 As for the Specific Test Question, we agree that the mere fact that the non-exhaustive factors listed in Thomas Heng were not applied does not give rise to a question of law. However, given that the overarching legal standard being articulated was, with respect, unclear, it stood to reason that how the “unusually convincing” standard was to be applied was thrown into doubt, and the Specific Test Question would therefore also arise as a question of law of public interest.

(2) Question of law arising as to the Inherent Reliability Question

74 We turn to consider the distinction drawn by the Judge between the reliability of the evidence of an alleged victim and that of an eyewitness. In his Oral Grounds, the Judge highlighted that:

This was elaborated further in his Written GD, where he stated that:

36 Nurse MJ was a bystander. She was not the victim. While an eye witness’s testimony has the advantage of being an independent account of the events, eye witness evidence is always subject to possible misapprehension, and errors in observation. In contrast to a victim’s own testimony, an eye witness’s account would be subject to a greater degree of misperception, misapprehension and misattribution. This is particularly so where the incident might have taken place over a period of time and the eye witness was only present for a brief moment. Without an appreciation of the full context in which the events unfolded, a bystander’s account may be liable to misinterpretation. An eye witness’s testimony is not a recording. …

…

38 … A key aspect of the District Judge’s decision was that Nurse MJ could not have been mistaken as to what she saw. However, as I have explained above, it must be appreciated that Nurse MJ’s perception of the matter would be weaker than the account of the [V]ictim. The incident might well have transpired. But having examined the existing state of the evidence as a whole, I did not find that the Prosecution had proven its case beyond reasonable doubt.

[emphasis added in italics and bold italics]

75 In our judgment, the above observations by the Judge similarly give rise to a question of law of public interest. We disagree with Mr Lau that the Judge was only laying down a “general starting proposition” and not an absolute statement. It seems to us that the Judge’s comment that an eyewitness’s observations would “always” be weaker than those of an alleged victim was a categorical statement of law. In any event, even if the Judge were only laying down a general starting proposition, this statement of law would be novel, and of significance for the evaluation of not just eyewitness testimony, but also the testimony of alleged victims. We are therefore satisfied that the third condition stated in Mohammad Faizal (see [64(c)] above) has been met with respect to the Inherent Reliability Question, and the Reframed Question more generally.

Whether the Judge’s determination of law had affected the outcome of the case

76 Mr Lau submitted that the fourth condition set out in Mohammad Faizal (see [64(d)] above) had not been met in any event, as the outcome of an acquittal should nonetheless remain in place for two reasons. First, the Prosecution’s grievances over the Judge’s decision were ultimately of a factual and evidentiary nature, and not of a legal nature, and the Judge was entitled to place more weight on certain factors. Second, regardless of how the “unusually convincing” standard was to apply in the assessment of an eyewitness’s evidence, the relevant factors or guidelines ought only to apply to future cases where identification was in issue.

(1) The ambit of the fourth condition set out in Mohammad Faizal

77 With respect, we are not persuaded by Mr Lau’s first submission. As the starting point of our analysis, we accept that construing s 397 of the CPC too freely to allow a criminal reference would “seriously undermine the system of one-tier appeal”, and that “[t]he interests of finality would strongly militate against the grant of … a [criminal] reference save in very limited circumstances” even if the question concerned were referred by the Public Prosecutor: see Mohammad Faizal at [21] and Lam Leng Hung at [50].

78 In Mohammad Faizal at [26]–[27], this court was of the view that the criminal reference brought by the applicants did not meet the fourth condition listed at [64(d)] above because no decision as a matter of “judgment or sentence” had been made by the trial court in relation to the applicants’ pending charges. Matters are quite different here, since the Respondent’s appeal in MA 9156 has resulted in the definitive outcome of an acquittal. Of course, the mere fact that an outcome has been reached is not dispositive of whether the fourth condition has been satisfied. As this court stressed in Mohammad Faizal at [27], the crucial word in that part of s 397(1) of the CPC which corresponds to the fourth condition is whether the High Court’s determination “affected” the outcome of the case.

79 In our judgment, to prevent an abuse of the criminal reference procedure, it is necessary that the High Court’s determination of the question of law significantly impacts, if not be dispositive of, the outcome of the case in order for the fourth condition in Mohammad Faizal to be satisfied. Were it otherwise, a question of law (though of public interest) could trigger the court’s substantive jurisdiction even though the question is ultimately academic in nature.

80 In the present case, the Judge’s articulation of the relevant legal standards and propositions indubitably resulted in the Respondent’s acquittal, and were in fact an intrinsic part of his reasoning leading to that outcome. This was not a situation where the Prosecution was simply alluding to some unsettled question of law in order to avail itself of a second tier of appeal.

(2) Prospective overruling

81 For completeness, we should state that we find ourselves unable to agree with Mr Lau’s second submission that any answer to the Specific Test Question (namely, how the “unusually convincing” standard should be applied) should only apply to future cases. As we indicated to Mr Lau at the hearing before us, the doctrine of prospective overruling was not applicable to the present case.

82 In Adri Anton Kalangie v Public Prosecutor [2018] 2 SLR 557 at [31], [33] and [40], this court clarified that the central precept of the doctrine of prospective overruling is to limit the temporal effect of a court’s judicial pronouncements. The exercise of the court’s discretion in this regard is guided by: (a) the extent to which the law or legal principle concerned is entrenched; (b) the extent of the change to the law; (c) the extent to which the change to the law is foreseeable; and (d) the extent of reliance on the law or legal principle concerned. No one factor is preponderant over any other, and no one factor must necessarily be established before the doctrine can be invoked in a particular case. The discretion to invoke this doctrine should only be exercised in exceptional circumstances where it is necessary to “avoid serious and demonstrable injustice to the parties or to the administration of justice”.

83 Traditionally, discussion of whether the doctrine of prospective overruling should be invoked has taken place in the context of sentencing guidelines: see Poh Boon Kiat v Public Prosecutor [2014] 4 SLR 892 at [113]; see also Chang Kar Meng v Public Prosecutor [2017] 2 SLR 68 at [75]. It has also been considered in civil cases, though we previously cautioned that the exceptionality of the doctrine being invoked would likely be even more prominent in the civil context: see Vinmar Overseas (Singapore) Pte Ltd v PTT International Trading Pte Ltd [2018] 2 SLR 1271 at [143].

84 In the present case, the Reframed Question concerns the Respondent’s criminal liability to a conviction for the offence of outrage of modesty under s 354(1) of the Penal Code generally, and the evidential rules pertaining to that offence specifically. This is therefore not a situation where exposure to the degree of punishment meted out is in issue. Rather, the factual question is whether or not the Respondent committed the acts alleged in the OM Charge. This in turn concerns the legal question of how Nurse MJ’s observations of those acts should be evaluated as a matter of evidence, and, more broadly, how such evidence should be weighed in terms of the burden of proof. To be clear, our foregoing remarks are not intended to exclude the doctrine of prospective overruling from applying in exceptional cases concerning criminal liability for offences or evidential rules. However, in the present case, it could not be said that there was any reliance by the Respondent on the questions of law raised in the present Criminal Reference.

85 In the circumstances, we are satisfied that the Reframed Question meets the fourth condition set out in Mohammad Faizal. Accordingly, we exercise our substantive jurisdiction to answer the Reframed Question.

Our answer to the Applicable Standard Question

86 We address, first, the Applicable Standard Question. The Prosecution submitted that the “unusually convincing” standard is to be applied when evaluating the uncorroborated evidence of an eyewitness to an offence, where such evidence forms the sole basis for a conviction. Mr Kow further submitted that, in principle, the “unusually convincing” standard would apply to non-sexual offences as well as sexual offences. As we highlighted earlier at [57] above, the Defence accepts both propositions.

The “unusually convincing” standard applies to all instances where the uncorroborated testimony of an eyewitness forms the sole basis for a conviction

87 We agree with the parties that the “unusually convincing” standard is the only standard to be applied where an eyewitness’s uncorroborated testimony forms the sole basis for a conviction. With respect, we are of the view that the Judge erred when he appeared to suggest that there were differing standards in relation to eyewitnesses as opposed to alleged victims.

88 The “unusually convincing” standard is used to describe a situation where the witness’s testimony is “so convincing that the Prosecution’s case [is] proven beyond reasonable doubt, solely on the basis of the evidence”: see Mohammed Liton at [38]. In Haliffie bin Mamat v Public Prosecutor and other appeals [2016] 5 SLR 636 at [28], this court considered that “a complainant’s testimony would be unusually convincing if the testimony, ‘when weighed against the overall backdrop of the available facts and circumstances, contains that ring of truth which leaves the court satisfied that no reasonable doubt exists in favour of the accused’” (citing Mohammed Liton at [39]). The relevant considerations in this regard include the witness’s demeanour, and the internal and external consistencies of the witness’s evidence.

89 In our judgment, the “unusually convincing” standard is necessarily applicable to the evidence of an eyewitness, just as it would apply to that of a complainant or an alleged victim, as long as the testimony of the witness in question is uncorroborated and therefore forms the sole basis for a conviction. This is because the basis for the “unusually convincing” standard has nothing to do with the status of the witness concerned (namely, whether he or she is an alleged victim or an eyewitness), and instead has everything to do with “the ultimate rule that the Prosecution must prove its case beyond a reasonable doubt”: see XP at [31]. In the absence of any other corroborative evidence, the testimony of a witness, whether an eyewitness or an alleged victim, becomes the keystone upon which the Prosecution’s entire case will rest. Such evidence can sustain a conviction only if it is “unusually convincing” and thereby capable of overcoming any concerns arising from the lack of corroboration and the fact that such evidence will typically be controverted by that of the accused person: see the decision of this court in AOF v Public Prosecutor [2012] 3 SLR 34 (“AOF”) at [111].

90 Put simply, the “unusually convincing” standard entails that the witness’s testimony alone is sufficient to prove the Prosecution’s case beyond a reasonable doubt: see Teo Keng Pong v Public Prosecutor [1996] 2 SLR(R) 890 at [73]. The overwhelming consideration that triggers the application of the standard is the amount and availability of evidence: see also Kwan Peng Hong at [29].

91 In a sense, the “unusually convincing” standard is not a “test” at all, but rather, a heuristic tool. It is a cautionary reminder to the court of the high threshold that the Prosecution must meet in order to secure a conviction, and of the anxious scrutiny that is required because of the severe consequences that will follow from a conviction. That scrutiny is directed towards the sufficiency of a witness’s testimony, which is inextricably linked to the ultimate inquiry of whether the case against the accused person has been proved by the Prosecution beyond a reasonable doubt: see Mohammed Liton at [39].

92 As we pointed out to Mr Kow during the oral hearing, because the “unusually convincing” standard is directed towards whether the Prosecution has satisfied its onerous burden of proof, it applies as a cautionary reminder at the last stage of the evaluation of the evidence and just before a conviction is found. The “unusually convincing” standard is not meant to impose a mandatory warning from the judge to himself or herself: see Mohammed Liton at [39]. Rather, its aim is to ensure that the trial judge has an awareness of the dangers of convicting the accused person on uncorroborated evidence, and that he or she (as well as an appellate court) undertakes a rigorous and holistic assessment of the evidence.

The abolition of the “suspect categories” of witnesses

93 Once it is appreciated that the “unusually convincing” standard is invoked in respect of uncorroborated evidence, the distinction drawn between the evidence of an eyewitness and that of an alleged victim is, with respect, unsustainable. We would eschew such a dichotomy for the additional reason that it may unintentionally lead to invidious distinctions. To appreciate why this is so, it is necessary to briefly consider how the common law historically treated the evidence of “suspect” witnesses such as alleged sexual offence victims and children. In cases involving such witnesses, the common law often prevented the jury from returning a conviction if it was solely predicated on the evidence of these witnesses, and demanded corroboration under the rule outlined in The King v Baskerville [1916] 2 KB 658 (“Baskerville”).

94 The rationale for this rule was the problematic notion that the evidence of such witnesses should be treated with suspicion. Prof Glanville Williams articulated this view in Corroboration – Sexual Cases [1962] Crim LR 662 as follows (at p 662):

On a charge of a sexual offence, such as rape, indecent assault or [a] homosexual offence, it is the practice to instruct the jury that it is unsafe to convict on the uncorroborated evidence of the alleged victim. There is sound reason for this, because sexual cases are particularly subject to the danger of deliberately false charges, resulting from sexual neurosis, phantasy, jealousy, spite, or simply a girl’s refusal to admit that she consented to an act of which she is now ashamed. Of these various possibilities, the most subtle are those connected with mental complexes. …

95 Such views were widely castigated, with one commentator referring to them as underpinned by “neanderthal ideas and attitudes”: see Ronald Joseph Delisle, Evidence: Principles and Problems (Carswell, 5th Ed, 1999) at p 428. Others considered that this need for corroboration ignored the reality that “many sexual offences are committed in circumstances in which corroboration is difficult if not impossible to obtain”: see Adrian Keane & Paul McKeown, The Modern Law of Evidence (Oxford University Press, 9th Ed, 2012) at p 227. Over time, the requirement for strict corroboration in sexual offences was abolished in England.

96 In our jurisprudence, the requirement for strict corroboration in the Baskerville sense has not been followed. Instead, our courts adopt a liberal approach to corroboration, focusing instead on the substance, relevance, and confirmatory value of the evidence in question: see AOF at [173]–[174]. As for the approach of treating the evidence of an alleged sexual victim with suspicion, Yong Pung How CJ made the following salutary remarks in Kwan Peng Hong at [27]:

27 … [I]t is objectionable to argue that extreme caution is required because female witnesses are prone to fantasising, exaggeration and lies due to some sexual neurosis. From my numerous judgments in the last ten years, I hope I have made clear that the court will no longer entertain such an argument, whatever the attitude was 20 years ago. Such generalised categorisation of female witnesses in sexual offence cases, based on the perceived dangers of false accusation caused by sexual neurosis, jealousy, fantasy, spite or shame, is not acceptable to a Singapore court today. Such dangers can easily be present in other cases too. To bring this up again in cases involving sexual offences, without pointing to specific facts justifying such extreme caution in a particular case, is both disingenuous and rather offensive in my view, and incongruous with the societal norm today. There would need to be an evidential basis for suggesting that the evidence of the witness might be unreliable. Mere suggestion by counsel would not be sufficient. [emphasis added]

97 We note in passing that in Canada, children are no longer regarded as a suspect category of witnesses. As the Supreme Court of Canada observed in R v W(R) [RW] [1992] 2 SCR 122 (“RW”) at [25]–[26], the evidence of children should not be assessed from the perspective of rigid stereotypes, but on a common sense basis. Each child witness, regardless of his or her age, should have his or her credibility and evidence assessed by reference to criteria appropriate to his or her mental development, understanding, and ability to communicate. We note further that the observations of McLachlin J (as she then was) in RW have in part been echoed by our High Court in B v Public Prosecutor [2003] 1 SLR(R) 400 at [25].

98 In our judgment, drawing a distinction between the evidential standards to be applied to an eyewitness on the one hand, and to an alleged sexual offence victim on the other, may unintentionally create a problematic dichotomy and reintroduce the antediluvian notions that were inherent in the suspect categories of witnesses. If an eyewitness were subject to a less stringent standard than the “unusually convincing” standard, it would implicitly suggest that sexual offence victims are inherently less honest than eyewitnesses, and that their evidence needs to be treated with more suspicion.

The “unusually convincing” standard applies across all types of offences

99 We also agree with Mr Kow that the Judge’s allusion at [33] of his Written GD that a different standard might apply to non-sexual offences ought to be clarified. As we highlighted at [52] above, the application of the “unusually convincing” standard does not depend on the category of offence alleged against the accused person, but rather, on the uncorroborated nature of the evidence presented to the court.

100 This view was expressed by Yong CJ in Kwan Peng Hong at [28] as follows:

28 … I am aware that in cases involving sexual offences, making an allegation is easy and rebutting it rather difficult. That is why … evidence of such an allegation must be sifted with care. But this should be done in all cases, where the court is faced with two contested versions of events and has to choose one, for a decision one way or the other … [emphasis added]

101 Similarly, in XP at [33], Rajah JA cautioned against applying a different standard to the evidence of witnesses from the suspect categories (citing Prof Michael Hor, “Corroboration: Rules and Discretion in the Search for Truth” [2000] Sing JLS 509 at p 518):

33 This reminder should not … be confined to the categories of witnesses who are supposedly accomplices, young children, or sexual offence complainants. Prof Michael Hor (Michael Hor, “Corroboration: Rules and Discretion in the Search for Truth [2000] Sing JLS 509) rightly observes (at 518) that the categorical approach is both under- and over-inclusive:

It is clear that witnesses of potentially doubtful credibility may fall outside of these categories: witnesses with a grudge against the accused, or witnesses who stand to gain something by incriminating the accused. … Conversely, there are witnesses who are within the classic corroboration categories, but whose credibility is not any more in question than any other witness …

102 In our judgment, an application of a different standard for eyewitnesses specific to sexual offences would ossify an unwarranted division between sexual offences and other offences. There is no reason to think that the independence of an eyewitness in a non-sexual offence would be any greater than that of an eyewitness in a sexual offence. Insofar as a motive for a false allegation is raised, in both sexual and non-sexual offence cases, it is for the Defence to first establish sufficient evidence of such a motive (see AOF at [215]). We would add that that motive must be specific to the witness concerned. General assertions without more would not ordinarily suffice. The foregoing reasons also compelled us at [52]–[53] above to reframe the Applicable Standard Question so as to exclude references to the independence of an eyewitness and to sexual offences, which are irrelevant to the matter at hand.

103 Before we conclude our answer to the Applicable Standard Question, we make one other observation. The Prosecution submitted that the “unusually convincing” standard should apply “[s]o long as it is easy for a witness to make an allegation but difficult for the accused to rebut it, and a court has to acquit or convict the accused solely based on the witness’s allegation”. This somewhat regrettable use of language appears to have been unintentionally cited from Kwan Peng Hong at [28] (see the passage reproduced at [100] above). Regardless of the veracity of an allegation of sexual assault, the reporting of the alleged assault should not be depicted as “easy”, particularly given the considerable practical, psychological, and emotional barriers that sexual assault victims face in bringing their violators to account: see Paul Roberts & Adrian Zuckerman, Criminal Evidence (Oxford University Press, 2nd Ed, 2010) at p 672. We consider that it is much more preferable to ground the “unusually convincing” standard solely on the basis that the witness’s evidence is uncorroborated. There is no need to invoke the language of whether allegations of sexual assault are “easy to make” or “difficult to refute”.

104 In the circumstances, we answer the Applicable Standard Question as such:

[Answer] The “unusually convincing” standard applies to the uncorroborated evidence of a witness in all offences (and not just sexual offences), where such evidence forms the sole basis for a conviction. In principle, the standard applies regardless of whether the witness is an eyewitness or an alleged victim.

Our answer to the Inherent Reliability Question

105 We turn to the Inherent Reliability Question. Mr Lau submitted that the Judge was not wrong to have held at [36] of the Written GD that an eyewitness’s observations would be “subject to a greater degree of misperception, misapprehension and misattribution” compared to a victim’s testimony. Mr Lau also submitted that in any event, this was not a “strict and immutable legal principle”, but merely a “general starting proposition” grounded in common sense. All things being equal, Mr Lau argued, an eyewitness’s observations would generally be less reliable. In this regard, he relied on this court’s observations in Sandz Solutions (Singapore) Pte Ltd and others v Strategic Worldwide Assets Ltd and others [2014] 3 SLR 562 (“Sandz Solutions”).

106 We find ourselves unable to agree with Mr Lau that the Judge, in making the aforesaid comment at [36] of his Written GD, was only laying down a general starting proposition. As we observed earlier at [75] above, the Judge appeared to have made a categorical statement of law. In particular, he also stated at [36] of his Written GD that “eye witness evidence is always subject to possible misapprehension and errors in observation” [emphasis added], and went on to elaborate at [38] that “it must be appreciated that Nurse MJ’s perception of the matter would be weaker than the account of the [V]ictim” [emphasis added].

107 Even if these statements by the Judge were only a starting proposition, we do not agree that that proposition is a correct statement of law for several reasons. First, as the Prosecution pointed out, the observations by this court in Sandz Solutions at [56] as to the fallibility of memory and perception were made in the context of witnesses in general and not specifically in relation to eyewitnesses (and certainly were not meant to draw a distinction between alleged victims and eyewitnesses).

108 Second, and more importantly, we are not persuaded that any general rule can be laid down about the relative evidential reliability of eyewitnesses compared to victims because this depends on all the circumstances of the case. An observation depends on several objective factors, among them, time, space, location, and line of sight. There are also subjective factors such as, on the one hand, possible intoxication, fatigue, impediments, intellectual defects, and, on the other, specialised training. Further, as elaborated in Sandz Solutions at [49]–[55], even after an observation is made, the process of memory recall and reconstruction is susceptible to error. The factors influencing the perception and recollection of a witness (whether an alleged victim or an eyewitness) permit of infinite variation and are not susceptible to cataloguing, much less generalisations.

109 Mr Kow helpfully pointed out that in some situations, the proximity of the victim to the offence does not necessarily aid the victim’s powers of observation. In Tan Wei Yi v Public Prosecutor [2005] 3 SLR(R) 471, for instance, the appellant was acquitted of a charge of voluntarily causing grievous hurt in furtherance of a common intention because the victim had never actually seen which of the three alleged assailants had actually punched him while he was being attacked (at [30]). In the same vein, there may be situations where an eyewitness might be better placed than a victim to make observations (such as if the eyewitness had received specialised training in observing details): see the comments of Kerr LCJ (as he then was) in the Northern Ireland Court of Appeal’s decision in R v Hagans [2004] NIJB 228 (“Hagans”) at [49]. We think that the observations of any witness, whether an eyewitness or an alleged victim, can be either compromised or, conversely, aided by a plethora of non-exhaustive factors that cannot be stated in advance. The crucial question in each case is whether there is evidence specific to that particular eyewitness or alleged victim which shows that the quality of his or her observations has been compromised or, conversely, improved. Again, general assertions without more will not suffice.

110 Aside from the objective circumstances and the subjective characteristics of the witness concerned, there is also a crucial question of what is being observed and recalled and its context. An observer watching a busy road for hours and asked to recall a single car may not fare better than another observer who observed the same road for a shorter duration. Repeated and mundane experiences are yet another case in point. Even individuals with perfectly functional memories would be hard-pressed to recall uneventful details such as a meal or the journey to work: see Gillian Cohen, “Human memory in the real world” in Analysing Witness Testimony: A Guide for Legal Professionals and Other Professionals (Anthony Heaton-Armstrong, Eric Shepherd & David Wolchover eds) (Blackstone Press Limited, 1999) ch 1 at p 12. All of the foregoing point to the dangers of drawing overly broad generalisations about the cogency of a particular category of witnesses’ observations divorced from the specific context in which the observations were made.

111 Third, we consider that holding that the evidence of eyewitnesses is weaker than that of alleged victims, particularly in the context of sexual offences, may have the converse and invidious consequence of placing a more onerous burden on victims of sexual assault to recollect the details of their assault. In other words, a sexual assault victim may instead be expected to have observations of greater clarity and consistency, since he or she is after all the person who experienced the sexual assault. Academic literature, on the other hand, shows that at the moment of sexual assault, a substantial number of victims may experience “tonic immobility”, which is an involuntary temporary state of inhibition. In other words, instead of the usual “fight or flight” response, some victims may “freeze” in intense fear: see Anna Möller, Hans Peter Söndergaard & Lotti Helström, “Tonic immobility during sexual assault – a common reaction predicting post-traumatic stress disorder and severe depression” (2017) 96 Acta Obstetricia et Gynecologica Scandinavica 932 at p 935. Tonic immobility affects memory processes, causing vivid memory recall in some victims, while causing others to “check out”: see Sunda Friedman TeBockhorst, Mary Sean O’Halloran & Blair N Nyline, “Tonic Immobility Among Survivors of Sexual Assault” (2015) Psychological Trauma: Theory, Research, Practice, and Policy 7(2) 171 at pp 173 and 176.

112 In this regard, courts have been advised to proceed with caution in making generalisations about observations and memory (see Mark L Howe, “The Neuroscience of Memory Development: Implications for Adults Recalling Childhood Experiences in the Courtroom” (2013) Nature Reviews Neuroscience 14(12) 869 at p 881):

… It is also important for jurors and judges to know that experiences that are encoded, stored, or retrieved during times of stress are not more likely to be remembered. Indeed, stress can actually impair the encoding and storage of autobiographical experiences and reduce the ability to retrieve specific episodic information during subsequent recall attempts.

These scientific findings stand in stark contrast to judges’ and jurors’ beliefs about memory and its development … This gap needs to be closed so that decisions about guilt or innocence in the courtroom reflect the scientific ‘truths’ about memory and not simply ‘common sense’ beliefs of judges and jurors. …

113 Similarly, the scientific literature also shows that an individual’s capacity for observation and memory recall may not always lie on a continuum even when the account in question concerns events occurring within the same episode (see James Hopper & David Lisak, “Why Rape and Trauma Survivors Have Fragmented and Incomplete Memories” (Time, 9 December 2014)):

… Victims may remember in exquisite detail what was happening just before and after they realized they were being attacked, including context and the sequence of events. However, they are likely to have very fragmented and incomplete memories for much of what happens after that.

These advances in our understanding of the impact of trauma on the brain have enormous implications for the criminal justice system. It is not reasonable to expect a trauma survivor – whether a rape victim, a police officer or a soldier – to recall traumatic events the way they would recall their wedding day. They will remember some aspects of the experience in exquisitely painful detail. Indeed, they may spend decades trying to forget them. They will remember other aspects not at all, or only in jumbled and confused fragments. Such is the nature of terrifying experiences, and it is a nature that we cannot ignore.

114 In our judgment, a general rule distinguishing between the reliability of an eyewitness’s account and that of an alleged victim’s account is unhelpful and may give rise to the dangers alluded to above. Accordingly, we answer the Inherent Reliability Question in the negative:

[Answer] No. The evidence of an eyewitness is neither less nor more reliable than that of an alleged victim. The reliability of any witness’s observation and account must be assessed in light of all the circumstances in each individual case.

Our answer to the Specific Test Question

Whether the Thomas Heng test for identification evidence should be adopted

115 We turn to the Specific Test Question. The Prosecution submitted that given that the “unusually convincing” standard should have been used to assess Nurse MJ’s evidence, the Judge erred in stating that her evidence “fell short of the required cogency and strength” needed to secure a conviction, and that her evidence of the Respondent’s alleged assault on the Victim was “not definitive and conclusive of the matter” (see the Written GD at [33] and [35]). According to Mr Kow, in a situation where the accused person has alleged that an eyewitness has mistakenly identified him or her as the perpetrator of an offence, the issue is to be addressed by the factors highlighted in Thomas Heng, which in turn form a subset within the “unusually convincing” standard.

116 In Thomas Heng, the appellant was convicted of a charge of trafficking in diamorphine. This court had to determine whether the Prosecution’s case was proved beyond a reasonable doubt on the basis of two eyewitnesses who had identified the appellant as the person who had placed a plastic bag containing the diamorphine into a dry riser compartment.

117 At [32] of Thomas Heng, this court considered the guidelines articulated by the English Court of Appeal in Regina v Turnbull and another [1977] QB 224 (“Turnbull”) in respect of the reliability of identification evidence. This court then adapted the Turnbull guidelines in the form of the following three-step test (at [33]–[35]):

(a) In a case involving identification evidence, the court should consider whether the case against the accused person depends wholly or substantially on the correctness of the identification evidence which is alleged by the Defence to be mistaken.

(b) If that is the case, the court should consider whether the identification evidence is of good quality, taking into account a non-exhaustive list of factors, among them, the length of time of the witness’s observation, the distance at which the observation was made, the presence of obstructions in the witness’s line of sight, the number of times the witness had seen the accused person previously, the frequency with which the witness had seen the accused person, the presence of any special reasons for the witness to remember the accused person, the length of time which elapsed between the original observation of the accused person and the subsequent identification of the accused person to the police, and the presence of any material discrepancies between the description of the accused person given by the witness and the actual appearance of the accused person. The court should take note of any specific weaknesses in the identification evidence. It can only safely assess the value of the identification evidence if it is satisfied that the quality of the identification is good.

(c) If the quality of the identification is poor, the court should consider whether there is any other evidence that can support the correctness of the identification. If the court is unable to find such other supporting evidence, it should be mindful that a conviction based on poor identification evidence would be unsafe.

118 We find ourselves hesitant to accept Mr Kow’s submission on the Thomas Heng test for several reasons. First, the Turnbull guidelines were formulated in the context of a trial judge’s summing-up directions to a jury (see Turnbull at 228). Although this court sought in Thomas Heng at [33]–[35] to rework the Turnbull guidelines to adapt them to our criminal trial system, we do not read the three-step test articulated in Thomas Heng as a prescriptive formula that must invariably be applied, but rather, as a set of possible pointers.

119 Second, as Mr Lau correctly pointed out, both Turnbull and Thomas Heng were considered in the context of identification evidence, and not simply in the context of all categories of eyewitness evidence. The former concerns recognition, whereas eyewitness evidence of the sort that we are concerned with concerns the witnessing of some activity. The Turnbull guidelines were designed to answer the question “who was there”, rather than the slightly different inquiry of “who did what”: see Andy Roberts, “Questions of ‘who was there?’ and ‘who did what?’: the application of Code D in cases of dispute as to participation but not presence” [2003] Crim LR 709 at p 712. In the present case, it is undisputed that the Respondent was in the proximity of Bed 7 (the Victim’s bed) at the material time. It is also undisputed that Nurse MJ had correctly identified the Respondent in her observations. What is in dispute was what the Respondent was doing to the Victim.

120 In this regard, we note that the nature of identification evidence is binary (the accused person either was or was not the person at the scene), whereas eyewitness evidence may be multi-factorial and concerned with details. A similar distinction was observed by the English Court of Appeal in Regina v Gary Shawn Linegar [2001] EWCA Crim 2404. There, Kay LJ commented, in respect of a situation where “[o]ne person did one thing, another person did a different thing”, that “[t]hose [were] not the circumstances to which the case of Turnbull and [the] related cases were directed at all” as “[t]here was no element of identification which was in issue” (see also the observations of the High Court of Auckland in Auckland City Council v Brailey [1988] 1 NZLR 103 at 106).

121 This subtle distinction can have significant consequences in terms of what a court should be minded to caution itself against. In Hagans at [48], Kerr LCJ made the following observations:

48 The need for a special warning in identification cases was acknowledged in R v Turnbull because of the experience of the courts that in such cases the potential for error is significant. Of course the danger is not only significant, it is also of a particular type. The nature of the warning that requires to be given in identification cases is therefore tailored to counteract the problems that identification throws up. Therefore the Court of Appeal in R v Turnbull referred to the need to warn the jury that witnesses, genuinely mistaken about a purported identification, may appear convincing; and that a close examination was required of the opportunity available to identifying witnesses to make a reliable identification. These warnings are considered necessary because of the tendency of some witnesses to feel certain that they have correctly registered the features of the individual that they believe that they can identify and because of the similarity in appearance of different people. Problems of a different nature arise where the identity of the participants in an incident is not in issue but what they did is. In that type of case, one is not concerned to guard against the potential for error in recalling accurately, for instance, the distinguishing personal characteristics of the person involved or in assessing whether the length of time available to observe the person identified was sufficient. Rather, the need for care in evaluating evidence of this kind arises because of the possibility of mistake about the roles played by the various protagonists where there is a fast moving, confused situation to be observed. It appears to us, therefore, that the warning to be given by a judge should be adapted to cater for the specific problem that is known to beset the particular species of evidence involved and that in the present case a classic Turnbull warning would not have been appropriate. [emphasis in original in italics; emphasis added in bold italics]

122 In other words, the dangers that the Turnbull guidelines were targeted at concerned a mistake in recognition arising from, among other things, an overconfidence in attributing identity despite a lack of prior acquaintance with or knowledge of the accused person’s appearance: see the decision of the Supreme Court of the Australian Capital Territory in Sharrett v Gill (1993) 113 FLR 316 at 323–324. Indeed, several of the factors referred to at the second stage of the Thomas Heng test (see [117(b)] above) were designed to mitigate exactly those problems because misidentification may stem from a witness guessing the accused person’s identity, or being driven to a positive identification by cues rather than by actual recognition. In contrast, the dangers to be guarded against in the present case may include, among other things, problems of perception, incorrect assumptions, or distorted recollection due to supplanted facts, each potentially attributing incorrect actions to an accused person or incorrect details to the incident in question.

123 Third, and inter-relatedly, this is not to say that the Thomas Heng test cannot be usefully adapted to eyewitness evidence. The Thomas Heng test is ultimately a method of analysing the evidence. Insofar as the factors in Thomas Heng are possible pointers, they form part of the forensic armoury that a judge has access to in assessing the evidence. There is no inexorable or inflexible rule that a judge must recite these factors in deciding whether to convict or acquit, as the case may be.

124 Most importantly, while we agree with Mr Kow that the Judge erred in his application of the “unusually convincing” standard, we are not persuaded that the error can narrowly be characterised as a purported failure to articulate and apply the Thomas Heng test to Nurse MJ’s evidence. In our judgment, the error lies a step prior in the reasoning as to the incidence of the burden of proof, and is of wider ambit than the factors enumerated in Thomas Heng.

125 As we see it, the error made by the Judge is not simply to do with whether the “unusually convincing” standard featured in this case. If this were so, the words “unusually convincing” would simply become an incantation. Mere recitation of this phrase cannot suffice to secure a conviction, nor does its omission mean that a reasonable doubt has arisen. The true question is how the “unusually convincing” standard has been applied, and, more importantly, whether the court has correctly carried out its task of determining whether the Prosecution has discharged its burden of proving its case against the accused person beyond a reasonable doubt.

The burden of proof beyond a reasonable doubt

126 The fundamental rule of proof beyond a reasonable doubt is considered hallowed precisely because it rests upon the bedrock principle of the presumption of innocence, which is the very foundation of criminal law. As a practical measure, the rule reduces the risk of convictions arising from factual error. This practical mechanism is itself grounded on the principle that allowing for the wrongful conviction of the innocent does violence to our societal values and fundamental sense of justice: see the concurring judgment of Harlan J in the United States Supreme Court case of In re Winship 397 US 358 (1970) (“Winship”) at 373, which was cited in Jagatheesan s/o Krishnasamy v Public Prosecutor [2006] 4 SLR(R) 45 (“Jagatheesan”) at [46]. But there is also an equally powerful rationale that animates the rule, which is that the coercive power of the State that flows from a conviction is legitimised precisely because it is based on this very principle of proof beyond a reasonable doubt. The faith that our society places in our criminal justice system stems from its confidence that only the guilty are punished: see the majority opinion delivered by Brennan J in Winship at 364; see also XP at [99], and Jagatheesan at [46] and [60].

127 As we alluded to at the start of this judgment, the phrase “beyond a reasonable doubt” is not susceptible to easy definition. Many have despaired of attempting to articulate its intricacies, and the English Court of Appeal was driven in R v Yap Chuan Ching (1976) 63 Cr App R 7 (“Ching”) to conclude at 11, somewhat cynically, that “if judges stopped trying to define that which is almost impossible to define there would be fewer appeals”. Despite these difficulties, we find ourselves attracted by the definition provided by Rajah JA in Jagatheesan at [55] that “[r]easonable doubt is, in other words, a reasoned doubt” [emphasis in original].

128 We note that the English Court of Appeal in Ching expressed a dislike for a similar formulation: see The Burden of Proof (Albert Kiralfy gen ed) (Professional Books Limited, 1987) at p 15. However, as Rajah JA observed in Jagatheesan at [57], the English Court of Appeal’s observations were made in the context of a jury system, where fact finders might not be legally trained. Such concerns, Rajah JA considered, were irrelevant in the context of Singapore, where fact finders were legally trained judges. We agree. In our judgment, the danger is not an inability to understand the concepts outlined, but that a proliferation of evidential rules might obscure the true task of the court, which is to scrutinise the evidence in light of the applicable standard of proof. In the subsequent analysis, we propose therefore to illustrate the operation of the principle of proof beyond a reasonable doubt, rather than seek to categorically prescribe its content.

The concepts underpinning the Prosecution’s burden of proof

129 The principle of proof beyond a reasonable doubt is simply that upon a consideration of all the evidence presented by the Prosecution and/or the Defence, the evidence must be sufficient to establish beyond a reasonable doubt each and every element of the offence with which the accused person is charged: see Jagatheesan at [48]. Inherent within the Prosecution’s weighty duty lies several concepts.

130 The first is the legal burden, which is the burden of proving a fact to the requisite standard of proof: see Jeffrey Pinsler, Evidence and the Litigation Process (LexisNexis, 6th Ed, 2017) (“Pinsler”) at para 12.007. This burden, which is encapsulated in ss 103 and 105 of the Evidence Act (Cap 97, 1997 Rev Ed) (“the Evidence Act”), is a burden that the Prosecution always bears in a criminal case: see Public Prosecutor v Mohd Ariffan bin Mohd Hassan [2019] 2 SLR 490 (“Mohd Ariffan”) at [112]. The obligation stems from the Prosecution’s responsibility to prove the charge against the accused person beyond a reasonable doubt, and it is “a permanent and enduring burden [which] does not shift” throughout the trial: see Britestone Pte Ltd v Smith & Associates Far East, Ltd [2007] 4 SLR(R) 855 at [60].

131 The second concept which underlies the Prosecution’s burden of proof is the degree of proof required. As Rajah JA noted in Jagatheesan at [52] (citing Wood JA’s judgment in the British Columbia Court of Appeal’s decision in R v Brydon (1995) 2 BCLR (3d) 243 at [82]–[83]), a quantitative description of the degree of proof would be circular and meaningless. Instead, what is required is a qualitative appreciation of whether a reasonable doubt has arisen. A reasonable doubt is one for which a reason can be given, so long as the reason given is logically connected to the evidence. As shall be discussed later in this judgment, a doubt is not reasonable if it is not reasoned, and the existence of a reasoned doubt is a necessary condition for an acquittal.

132 The third concept is the evidential burden, which is the burden to adduce sufficient evidence to raise an issue for the consideration of the trier of fact. Tactically speaking, the evidential burden can shift to the opposing party once it has been discharged by the proponent: see Public Prosecutor v BPK [2018] SGHC 34 (“BPK”) at [144]–[145]. The opposing party must then call evidence, or take the consequences, which may or may not be adverse: see Anti-Corrosion Pte Ltd v Berger Paints Singapore Pte Ltd and another appeal [2012] 1 SLR 427 at [37].

133 The evidential burden generally lies on the Prosecution, which has to “satisf[y] its evidential burden on [the] issue by adducing sufficient evidence, which if believed, is capable of establishing the issue beyond reasonable doubt”: see Colin Tapper, Cross and Tapper on Evidence (Oxford University Press, 12th Ed, 2010) (“Cross and Tapper”) at p 122. However, the evidential burden may also lie on the Defence, depending on the nature of the defence and the fact in issue that is being raised: see BPK at [146]. Regardless of the incidence of the evidential burden of proof, when a particular fact or defence raised by the accused person has properly come into issue, the Prosecution must rebut that fact or defence so as to meet its legal burden of proving the charge against the accused person beyond a reasonable doubt: see Cross and Tapper at p 122.

134 In our judgment, the principle of proof beyond a reasonable doubt can also be usefully conceptualised in two ways. First, a reasonable doubt may arise from within the case mounted by the Prosecution. To be clear, the term “within the case mounted by the Prosecution” should not be confused with the term “at the close of the Prosecution’s case”. The latter was articulated by the Privy Counsel in Haw Tua Tau and others v Public Prosecutor [1981–1982] SLR(R) 133 (“Haw Tua Tau”), and is now statutorily codified in s 230(j) of the CPC. It involves the procedural task of calling upon the accused person to give his defence. This takes place when the court is satisfied that there is some evidence which is not inherently incredible that satisfies every element of the charge. On the other hand, the former term proof beyond a reasonable doubt “within the case mounted by the Prosecution” denotes the evaluative task of considering all of the evidence adduced by the Prosecution at each stage of the proceedings.

135 Second, a reasonable doubt may arise on the totality of the evidence. As we shall explain further in this judgment, the totality of the evidence necessarily includes a holistic assessment of both the Prosecution’s and the Defence’s cases, and the interactions between the two. We now elaborate on these two perspectives.

Proof beyond a reasonable doubt within the Prosecution’s case

136 As we recently explained in Mohd Ariffan at [113], given that the legal burden lies on the Prosecution throughout a trial, as part of its own case, the Prosecution must adduce sufficient evidence to establish the accused person’s guilt beyond a reasonable doubt on at least a prima facie basis. One example of a failure to do so would be where, after the Defence has been called, there are discrepancies in the accused person’s testimony, but there remain significant inconsistencies in the Prosecution’s case that nevertheless generate a reasonable doubt. In such a situation, the court would be obliged to acquit the accused person. Another example would be where the Prosecution’s evidence is so weak that, at the close of the Prosecution’s case, it falls below the Haw Tua Tau standard. The court would then be entitled to find that there is no case to answer even without calling upon the Defence.

137 We must stress that in either of these situations, the court’s task remains to properly articulate the reasonable doubt that has arisen within the Prosecution’s case. In other words, the judge must be able to particularise the specific weakness in the Prosecution’s own evidence that irrevocably lowers it below the threshold of proof beyond a reasonable doubt. Such particularisation may include, among other things, an assessment of the internal consistency within the content of a witness’s testimony. It may also involve an assessment of the external consistency between a witness’s evidence and the extrinsic evidence, which includes testing the former against the inherent probabilities and uncontroverted facts: see Sandz Solutions at [39]; see also the observations of this court in Thorben Langvad Linneberg v Leong Mei Kuen [2013] 1 SLR 207 at [13]–[14].

138 Our reasoning finds expression in the following observations of Rajah JA in Jagatheesan at [55]–[56]:

55 … [T]he Court of Appeal has accepted that a reasonable doubt is one which is capable of distinct expression and articulation and has support and foundation in the evidence submitted which in the circumstances is essential to a conviction. As Prof Larry Laudan puts it, “What distinguishes a rational doubt from an irrational one is that the former reacts to a weakness in the case offered by the prosecution, while the latter does not”: see, Larry Laudan, “Is Reasonable Doubt Reasonable?” (2003) 9 Legal Theory 295 (“Larry Laudan”) at 320. Reasonable doubt is, in other words, a reasoned doubt.

56 … [T]his particular formulation of reasonable doubt … correctly shifts the focus from what could potentially be a purely subjective call on the part of the trial judge to a more objective one of requiring the trial judge to “[reason] through the evidence”: Larry Laudan at 319. Therefore, it is not sufficient for the trial judge merely to state whether he has been satisfied beyond reasonable doubt. He must be able to say precisely why and how the evidence supports the Prosecution’s theory of the accused’s guilt. This process of reasoning is important not only because it constrains the subjectivity of the trial judge’s fact-finding mission; it is crucial because the trial process should also seek to “persuade the person whose conduct is under scrutiny of the truth and justice of its conclusions”: R A Duff, Trials and Punishment (Cambridge University Press, 1986) at p 116; T R S Allan, Constitutional Justice: A Liberal Theory of the Rule of Law (Oxford University Press, 2001) at p 81.

[emphasis in original in italics; emphasis added in bold italics]

139 Although Rajah JA made his observations in the context of the reasons that a judge must supply to support a conviction, we find his observations equally apposite if a judge should instead decide to acquit. A judge must therefore supply his or her precise reasons for the acquittal when a reasonable doubt has arisen within the Prosecution’s case. This proposition can be illustrated by the case of Mohd Ariffan at [90]–[92] and [111], where this court was of the view that the discrepancies within the Prosecution’s case in terms of the evidence of the complainant, her mother, and a Prosecution witness (one Mr Sim) created a reasonable doubt in the Prosecution’s case which was not ameliorated by any corroborative evidence. There, we pointed out at [109] that Mr Sim’s evidence about the alleged scene of the crime directly contradicted the evidence of the complainant and her mother. This in turn impacted the complainant’s account of how frequently the alleged sexual assaults had occurred. Given these material discrepancies within the Prosecution’s case, the court reasoned at [113] that the mere fact that weaknesses in the accused person’s defence might support some aspects of the Prosecution’s case was not enough to secure a conviction.

140 Once the court has identified the weakness internal to the Prosecution’s evidence, in the absence of corroborative evidence that can militate against this weakness such that the Prosecution’s evidence as a whole can still prove the case against the accused person beyond a reasonable doubt, weaknesses in the Defence’s case cannot ordinarily shore up what is lacking in the Prosecution’s case to begin with (see [139] above). For instance, the adverse inference that may be drawn from an accused person’s silence is drawn precisely because the evidence adduced in the Prosecution’s case calls for an explanation that only the accused person can give. In other words, the Prosecution has already discharged its evidential burden, which has then shifted to the Defence. As observed in the majority judgment of this court in Took Leng How v Public Prosecutor [2006] 2 SLR(R) 70 at [43] (citing in part the decision of the High Court of Australia in Weissensteiner v R (1993) 178 CLR 217):

… A court would be in grave error if it were to draw an adverse inference of guilt if such an inference were used solely to bolster a weak case. … [T]he silence of the accused “cannot fill in any gaps in the prosecution’s case; it cannot be used as a make-weight”. …

141 The corroborative effect of lies told by an accused person (known as “Lucas lies” after the eponymous case of Regina v Lucas (Ruth) [1981] 1 QB 720 (“Lucas”)) serves as another useful illustration. In Lucas, the English Court of Appeal was satisfied that an accused person’s lies could corroborate other evidence against him, but only under carefully prescribed conditions: see Public Prosecutor v Ilechukwu Uchechukwu Chukwudi [2015] SGCA 33 at [60]. It is important to note that such lies can amount to corroboration of the accused person’s guilt if they are shown to have been told out of a motive that can only be linked to his guilt: see Er Joo Nguang and another v Public Prosecutor [2000] 1 SLR(R) 756 at [54]. Moreover, even when Lucas lies are established, they are only “employed to support other evidence adduced by the Prosecution; they cannot however by themselves make out the Prosecution’s case” [emphasis added]: see Public Prosecutor v Chee Cheong Hin Constance [2006] 2 SLR(R) 24 at [92].

142 These situations illustrate that in relation to the first way in which a reasonable doubt may arise, when attention is directed to examining the evidence within the Prosecution’s case, the court should consider whether the Prosecution’s evidence on its own is sufficient to meet the standard of proof beyond a reasonable doubt; and if it is not, the court should identify what it considers to be the inherent and irredeemable weaknesses in the Prosecution’s case. Once such flaws in the Prosecution’s case have been identified, weaknesses in the Defence’s case would not ordinarily operate to bolster the Prosecution’s case because the Prosecution has simply not been able to discharge its burden of proof beyond a reasonable doubt. The task of the court is to identify the relevant flaws within the Prosecution’s case.

Proof beyond a reasonable doubt on the totality of the evidence

143 We turn to the second way in which a reasonable doubt may arise, which is upon an assessment of the totality of the evidence. We consider the inquiry here to be intimately connected with the “unusually convincing” standard. As we indicated earlier (at [124] above), the Judge’s erroneous application of this standard is of wider import than simply an omission to apply the Thomas Heng test to Nurse MJ’s evidence. In this regard, it must be appreciated that the Prosecution’s case hinged on Nurse MJ’s evidence that she saw the Respondent straddling the Victim and sexually assaulting her. Nurse MJ’s account of the alleged sexual assault was uncorroborated by any objective evidence. Contrary to Nurse MJ’s account, the Respondent’s case was that he was present in the Room at the material time, but that he was instead assisting the Victim by reaching across her for her pillow and adjusting the pillow under her head. These were starkly different accounts. In our respectful judgment, the Judge’s misapplication of the “unusually convincing” standard lay in his failure to appreciate this crucial context of mutually exclusive and competing testimonies. As Rajah JA highlighted in XP at [31], “the ‘unusually convincing’ standard sets the threshold for the [witness’s] testimony to be preferred over the accused’s evidence where it is a case that boils down to one person’s word against another’s”.

144 The assessment of the Prosecution’s evidence under the “unusually convincing” standard must be made with regard to the totality of the evidence: see XP at [30]. The totality of the evidence logically includes the Defence’s case (both as a matter of the assertions put forth by the accused person, and the evidence he has adduced). The evaluative task here is not just internal to the Prosecution’s case, but rather, also comparative in nature. Where the evidential burden lies on the Defence and this has not been discharged, the court may find that the Prosecution has discharged its burden of proving its case beyond a reasonable doubt: see Pinsler at para 12.009, and Cross and Tapper at p 123. At this stage of the inquiry, regard may be had to weaknesses in the case mounted by the Defence as part of the assessment of the totality of the evidence.

145 Conversely, what the Defence needs to do to bring the Prosecution’s case below the requisite threshold is to point to such evidence that is capable of generating a reasonable doubt: see Pinsler at para 12.009. If the Prosecution fails to rebut such evidence, it will necessarily fail in its overall burden of proving the charge against the accused person beyond a reasonable doubt. We would add that such evidence need not necessarily be raised (in the sense of being asserted, or being made the subject of submissions) by the Defence in order for it to give rise to a reasonable doubt. What matters is that a reasonable doubt arises (in whatever form) from the state of the evidence at the close of the trial.

146 By way of illustration, in Mui Jia Jun v Public Prosecutor [2018] 2 SLR 1087 (“Mui Jia Jun”), the Prosecution’s case depended substantially on the co-accused person’s account that the appellant had handed him the drugs pre-packed within a bag. We held at [67(a)] that once the co-accused person’s DNA was found on the adhesive tape around the bundle of drugs, the evidential burden shifted to the Prosecution to explain this, and its failure to address this piece of evidence created a reasonable doubt as to whether the appellant had truly handed the co-accused a pre-packed bag of drugs, which was a central feature of the Prosecution’s case against the appellant, or whether the co-accused had himself handled the drugs. Mui Jia Jun demonstrates that whether a reasonable doubt has arisen must be particularised and duly identified in a manner that is specific to the Prosecution’s case: see [137] above.

147 As Rajah JA stated in Jagatheesan at [55] (citing Larry Laudan, “Is Reasonable Doubt Reasonable?” (2003) 9 Legal Theory 295 at p 320), “[w]hat distinguishes a rational doubt from an irrational one is that the former reacts to a weakness in the case offered by the prosecution, while the latter does not”. Similarly, in XP at [93], Rajah JA considered that a reasonable doubt “must react to a weakness in the case offered by the Prosecution”. Therefore, in order to make a finding that there is a reasonable doubt in the Prosecution’s case, the judge must articulate the doubt that has arisen in the Prosecution’s case on the totality of the evidence and then ground that doubt with reference to the evidence.

148 In the context of the uncorroborated evidence of an eyewitness, whether his or her account is considered unusually convincing (and therefore capable of discharging the Prosecution’s burden of proving the case against the accused person beyond a reasonable doubt) requires an assessment of the internal and external consistencies of the account, and of any other evidence that the court is bound to consider. Such other evidence necessarily requires a consideration of the Defence’s case and the evidence adduced by the accused person (or the lack thereof). We would also add that a finding must be made as to the relevant facts before the court directs itself to the ultimate inquiry of whether the Prosecution’s case has been proved beyond a reasonable doubt.

Summary of our answer to the Specific Test Question

149 The foregoing principles can be distilled into the following summary as our answer to the Specific Test Question:

(a) The factors in Thomas Heng serve as possible pointers and form part of the forensic armoury that a judge has access to in his or her assessment of the evidence. There is no inexorable or inflexible rule that he or she must recite them in deciding whether to convict or acquit, as the case may be.

(b) The principle of proof beyond a reasonable doubt entails that upon considering all the evidence presented by the parties, the evidence suffices to establish beyond a reasonable doubt each and every element of the charge against the accused person.

(c) The Prosecution’s legal burden to prove the charge against the accused person beyond a reasonable doubt does not shift throughout the proceedings. The term “beyond a reasonable doubt” requires a qualitative appreciation of whether a reasonable doubt has arisen, in the sense of a doubt that is supported by reasons that are logically connected to the evidence. A reasonable doubt is, in other words, a reasoned doubt, and is a necessary condition for an acquittal.

(d) Depending on the fact in issue and the nature of the defence, the evidential burden may lie on the Prosecution or on the Defence. Regardless of the incidence of the evidential burden, where a particular fact or defence raised by the accused person has properly come into issue, the Prosecution must rebut that fact or defence so as meet its overall legal burden of proving the charge against the accused person beyond a reasonable doubt.

(e) The principle of proof beyond a reasonable doubt can be conceptualised in two ways. First, a reasonable doubt may arise from within the case mounted by the Prosecution. As part of its own case, the Prosecution must adduce sufficient evidence to establish the accused person’s guilt beyond a reasonable doubt on at least a prima facie basis. Failure to do so may lead to a finding that the Prosecution has failed to mount a case to answer, or to an acquittal. In those situations, the court must nevertheless particularise the specific weakness in the Prosecution’s own evidence that irrevocably lowers it below the threshold of proof beyond a reasonable doubt.

(f) Once the court has identified the flaw internal to the Prosecution’s case, weaknesses in the Defence’s case cannot ordinarily shore up what is lacking in the Prosecution’s case to begin with, because the Prosecution has simply not been able to discharge its overall legal burden.

(g) The second way in which a reasonable doubt may arise is on an assessment of the totality of the evidence. The inquiry here is intimately connected with the “unusually convincing” standard, which arises in the context of mutually exclusive and competing testimonies. The “unusually convincing” standard sets the threshold for a witness’s testimony to be preferred over the evidence put forth by the accused person where it is a case of one person’s word against another’s.

(h) The assessment of the Prosecution’s evidence under the “unusually convincing” standard must be made with regard to the totality of the evidence, which logically includes the case mounted by the Defence. The evaluative task is not just internal to the Prosecution’s case, but also comparative in nature. Where the evidential burden lies on the Defence and this has not been discharged, the court may find that the Prosecution has discharged its burden of proving its case against the accused person beyond a reasonable doubt. At this stage, regard may be had to weaknesses in the Defence’s case.

(i) What the Defence needs to do to bring the Prosecution’s case below the requisite threshold is to point to such evidence as would generate a reasonable doubt. That evidence need not necessarily be raised by the Defence; what matters is that a reasonable doubt arises from the state of the evidence at the close of the trial.

(j) Whether a reasonable doubt has arisen must be particularised and duly identified in a manner that is specific to the Prosecution’s case. In order to make a finding that there is a reasonable doubt in the Prosecution’s case, the court must not only articulate the doubt that has arisen in the Prosecution’s case on the totality of the evidence, but also ground that doubt with reference to the evidence.

(k) Whether an eyewitness’s uncorroborated evidence is unusually convincing (and therefore capable of discharging the Prosecution’s burden of proving the case against the accused person beyond a reasonable doubt) requires an assessment of the internal and external consistencies of the eyewitness’s account, and of any other evidence that the court is bound to consider, which includes the Defence’s case and the evidence adduced by the accused person (or the lack thereof). A finding must be made as to the relevant facts before the court directs itself to the ultimate inquiry of whether the Prosecution’s case has been proved beyond a reasonable doubt.

Analysis of the facts in light of our answers to the Reframed Question

150 The crux of the Judge’s reasoning can be found at [36] and [38]–[39] of the Written GD, which we set out below:

…

38 The question in the present case is whether Nurse MJ’s testimony was sufficient to safely convict the [Respondent], bearing in mind the matters above and the absence of strong corroborating evidence. A key aspect of the District Judge’s decision was that Nurse MJ could not have been mistaken as to what she saw. However, as I have explained above, it must be appreciated that Nurse MJ’s perception of the matter would be weaker than the account of the [V]ictim. The incident might well have transpired. But having examined the existing state of the evidence as a whole, I did not find that the Prosecution had proven its case beyond a reasonable doubt.

[emphasis added]

151 As we see it, the Judge’s reasoning can be further narrowed down to his view that there was a possibility of a “mistake”, “misperception”, “misapprehension” or “misattribution” on Nurse MJ’s part: see the Written GD at [36], [38] and [39]. Given that the Judge erred in his articulation of the law in respect of the Applicable Standard Question and the Inherent Reliability Question, it is necessary to consider what the correct analysis should have been as to the assessment of the evidence in light of the principles elucidated above.

The Inherent Reliability Question on the facts

152 In our judgment, the Judge’s error in distinguishing between the reliability of an eyewitness’s account and that of an alleged victim’s account had a material impact on his decision. For one, his proposition that “eye witness evidence is always subject to possible misapprehension and errors in observation” (see the Written GD at [36]) was made about eyewitnesses in general and in the abstract. As we stated at [109] above, any finding of weaknesses in a witness’s observations must be specific to that witness.

153 In the present case, there was no finding that Nurse MJ’s line of sight was, for instance, obscured. Nor was it even suggested that she was too far away to pick out the details that she had alleged. On the contrary, it was undisputed that at the material time, Nurse MJ was positioned at the foot of Mdm MG’s bed (which was directly across from the Victim’s bed), standing about one-and-a-half arm’s lengths away from the Victim’s bed, and her view was unobstructed as the curtains were open across the width of the Victim’s bed: see [12]–[14] above.

154 More significantly, the distinction drawn by the Judge between eyewitnesses and alleged victims is erroneous not only because it is without foundation, but also because it is, in the present case, irrelevant. The Victim here was never in a position to testify in the first place, so, with respect, any comparison preferring the Victim’s account over Nurse MJ’s account misses the point as it was a comparison with a scenario that had never and could never have arisen.

The Specific Test Question on the facts

No reasonable doubt had arisen within the Prosecution’s case

155 At [44] of his Written GD, the Judge stated:

44 The inconsistencies in the evidence of the [Respondent] and [Mdm] JP did not go towards supporting or corroborating Nurse MJ’s evidence. The weakness of a defence does not in and of itself translate into [a] strength [in] the Prosecution’s evidence. It bears mentioning that the court does not have to believe an accused’s evidence to acquit the accused …

156 As a matter of principle, there is nothing objectionable in the Judge’s remarks, which are in line with our earlier observations. However, in our respectful view, the Judge erred in failing to identify and particularise the flaw within the Prosecution’s case that led him to believe that a reasonable doubt had arisen.

157 First, far from making a finding that Nurse MJ’s credibility was in issue, the Judge was of the view at [29] of his Written GD that there was “no reason to doubt the District Judge’s conclusion that Nurse MJ was honest”. In a similar vein, we reject Mr Lau’s submission before us that Nurse MJ had a motive to lie or embellish the facts because she “may have borne a grudge against [the Respondent]”. This submission was unsupported by evidence and devoid of credibility.

158 Second, and relatedly, it was not the Judge’s finding that what Nurse MJ had observed was impossible on the facts that she had narrated at the trial. In other words, there was nothing inconsistent or implausible within Nurse MJ’s testimony. On the contrary, given that Nurse MJ’s credibility was not in issue, there was no reason to doubt, among other things, her alleged position at the material time and her line of sight (see [153] above). Nor was there any reason to disbelieve her testimony on the daytime lighting conditions in the Room, her period of observation, and the fact that she was not fatigued when she made her observations.

159 Third, the potential impediment within the Prosecution’s case that the Judge did focus on at [39] of the Written GD was that “Nurse MJ had only about a five-second glimpse of the alleged assault”. In his Oral Grounds at [4(a)], the Judge also stated that “[Nurse MJ’s] viewing of the incident was perhaps not fleeting, but neither was it sufficient to be definitive”. With respect, nothing was identified about the purportedly short period of Nurse MJ’s observation that made the facts or the details alleged in her account insufficient or unreliable. Nor does it seem to us that the Judge was implying that Nurse MJ could not possibly have seen within that five-second window that which she described. Moreover, as we pointed out at [110] above, the duration of a witness’s observation must be viewed in the context of the observation. Nurse MJ was not describing some mundane occurrence, but the dramatic one of the Respondent straddling the Victim with his buttocks exposed and his groin in contact with the Victim’s exposed groin.

No reasonable doubt had arisen on the totality of the evidence

160 We are also of the respectful view that the Judge erred in his assessment of whether a reasonable doubt had arisen on the totality of the evidence and in his application of the “unusually convincing” standard. First, in his Written GD at [33]–[39], the Judge did not particularise what the purported mistake in Nurse MJ’s observations was. In our judgment, it was not sufficient for the Judge to find that some general “mistake” might have been made by Nurse MJ. Rather, it was necessary for him to state what the mistake made might be in relation to her account, so that he could then evaluate its likelihood.

161 Second, even where a doubt in the Prosecution’s case has been particularised, it is still necessary to ground it in the state of the existing evidence. In this regard, the Judge alluded peripherally that on the Respondent’s version, “he was attempting to adjust the [V]ictim’s head onto the pillow and reach for another pillow to support her head. It was consistent with this version that his posture might have been misperceived”: see the Written GD at [46]. However, a careful scrutiny of the facts does not disclose the possibility of such a “misperception”, and the Judge inadvertently omitted to assess the Respondent’s allegation of Nurse MJ’s “mistake” against the totality of the evidence. As the District Judge noted in the DJ’s GD at [53], Nurse MJ’s account of the Respondent straddling the Victim on top of her bed with his buttocks exposed was so drastically different from the Respondent’s account of his standing by her bed and assisting her with her pillow that it could not be explained as a mistake. It was ultimately a question of which of two wholly incompatible and mutually exclusive accounts was to be believed. Once the following factors are appreciated, we are satisfied that the Judge’s order of acquittal cannot be supported.

(1) The timing of the Respondent’s purported assistance to the Victim

162 In the course of his evidence, the Respondent narrated the following series of events:

(a) The CCTV footage showed that at 3.30.58pm, the Respondent entered the Room. The Respondent claimed that he began fixing the fuse of Mdm JP’s portable television at her bed, Bed 8. As he knelt down to switch on the power for Mdm JP’s television, he heard a sound from the Victim.

(b) The Respondent saw the Victim’s head tapping on the side railing of her bed (Bed 7) and thought she was in pain. He placed his left knee between the bars of the side railing to reach for the Victim’s round pillow, which he noticed was displaced. After tending to the Victim, he returned to adjusting the channels on Mdm JP’s portable television.

(c) While adjusting the channels on Mdm JP’s portable television, the Respondent noticed some “food greens” on Bed 8. The CCTV footage showed that the Respondent emerged from the inner section of the Room at 3.36.18pm. The Respondent claimed that he was, at that moment, heading to the toilet to dispose of the “food greens”.

(d) At 3.36.38pm, the Respondent returned to Bed 8, where he claimed to have continued working on Mdm JP’s portable television. The Respondent agreed that Ms SBR entered the Room and proceeded to Bed 8 to retrieve Mdm JP’s spectacles before exiting. The CCTV footage showed that Ms SBR came by sometime between 3.37.38pm and 3.38.02pm.

(e) From 3.38.02pm to 3.41.32pm (some three minutes and 30 seconds), no one was in the inner section of the Room, save for the Respondent and the Victim (as well as Mdm MG in Bed 6).

(f) At 3.41.32pm, Nurse MJ was seen entering the Room and exiting some 11 seconds later.

163 The Prosecution submitted that [162(e)] above showed that the Respondent remained in the Room for some three minutes and 30 seconds, during which he had uninterrupted access to the Victim. Moreover, the Respondent proffered contradictory explanations as to what he was doing in the Room after he had disposed of the “food greens” and repaired Mdm JP’s portable television, first accepting that the television was already repaired, and then claiming that he was fixing the plug to a power cable.

164 While the Prosecution’s submissions as to the duration for which the Respondent remained in the Room are significant, there is an even more glaring inconsistency in the Respondent’s defence as to the sequence of events. A key aspect of the Respondent’s defence was that Nurse MJ had misperceived his posture while he was assisting the Victim with her pillow. However, it is obvious that the Respondent could not have been assisting the Victim with her pillow when Nurse MJ entered the Room because there were at least two events (corroborated by the CCTV footage) that, on his case, had taken place after he had allegedly helped the Victim and before Nurse MJ entered the Room – his departure to the toilet to dispose of the “food greens”, and Ms SBR’s retrieval of Mdm JP’s spectacles. Based on the CCTV footage and the Respondent’s own narrative as summarised at [162(b)]–[(162(d)] above, and having regard to when Nurse MJ was seen entering the Room, the Respondent would already have reached for the Victim’s pillow and assisted the Victim at least five minutes before Nurse MJ even entered the Room. This was confirmed repeatedly with the Respondent during cross-examination and re-examination, and there was to his mind “[n]o confusion” about the sequence of events.

165 In our judgment, the Judge might not have appreciated that the Respondent’s allegation as to Nurse MJ’s misperception of his posture while he was assisting the Victim with her pillow was not a defence that pointed to a weakness in the Prosecution’s case. Put simply, since the Respondent could not (by his own account) have been assisting the Victim with her pillow when Nurse MJ saw him, there was nothing for Nurse MJ to be mistaken about. In order for a mistake to arise, there must be some objective reality to be misperceived, misattributed, or misinterpreted.

166 The aforesaid error might have been compounded by the Judge’s failure to particularise the purported mistake in Nurse MJ’s observation. In our judgment, there was no room for ambiguity or misperception in Nurse MJ’s account. On the contrary, the contemporaneous and almost immediate evidence arising from Nurse DS’s testimony was that Nurse MJ had asked him to “[p]lease go and see what [the Respondent] is doing on [the Victim’s] bed” [emphasis added]. The foregoing leads us to conclude that the Respondent’s own evidence had failed to cast a reasonable doubt (by way of a mistake in Nurse MJ’s observations) on the Prosecution’s evidence.

(2) The plausibility or otherwise of the Respondent’s posture while reaching for the Victim’s pillow

167 A similar error might have been made by the Judge in respect of the Respondent’s account of the position of his knees while he was reaching for the Victim’s pillow. As we observed at [156]–[159] above, no reasonable doubt had arisen internal to the Prosecution’s case. The inquiry therefore turned to whether the totality of the evidence, including the Defence’s case, could raise a particularised doubt to dispel the safety of a conviction and whether the Prosecution had successfully rebutted that reasoned doubt.

168 In this case, the Judge appeared to have suggested that the Respondent’s posture could have been misperceived and that the position of his knees was ultimately irrelevant (see the Written GD at [44]–[46]). This was a doubt that arose only as a result of the Defence’s evidence, and therefore, due consideration had to be given to whether the Prosecution had failed to rebut that doubt, and if so, why the Prosecution had failed in that regard; further, that particularised doubt had to be grounded in the evidence. In our judgment, the Prosecution had clearly rebutted that doubt by showing that the Respondent’s defence was externally and internally inconsistent.

169 First, we note the District Judge’s finding that regardless of whether the Respondent had placed one or both of his knees on the Victim’s bed while he was reaching for her pillow, his alleged position was “highly unnatural and contrived” as he could have easily reached for the pillow without placing either of his knees on the bed, given his height. In this regard, the District Judge had the benefit of a demonstration during the Trial Scene Visit, which was captured in the photographs adduced as Exhibits P91 and P94: see the DJ’s GD at [63]; see also [34] above. There was therefore absolutely no need for the Respondent to have placed either of his knees on the Victim’s bed while he was reaching for her pillow, rendering it even less probable that his alleged posture could have been misconstrued by Nurse MJ.

170 Second, the Respondent could not explain the inconsistencies in his statements to the police as to whether he had placed one or both of his knees on the Victim’s bed while he was reaching for her pillow. He claimed that he had corrected the reference in his First Statement to both his knees after having spoken to his fellow remandees: see [28] above. However, there is no evidence that he actually made such a correction in his Second Statement, save for a singular reference therein to “my knee”. ASP Razali, on the other hand, testified that the reference to “my knee” in the Second Statement was a typographical error: see [29] above. It seems to us that ASP Razali’s claim is supported by the weight of the evidence, given the other parts of the Second Statement where the Respondent did not correct erroneous references to “both” his knees.

171 Third, leaving aside the Respondent’s statements, during the Scene Investigation, which occurred just a day after the Respondent’s Second Statement was recorded, the Respondent told ASP Razali to place two arrows on Bed 7, corresponding to both his knees: see [30] above. The Respondent could not provide an explanation for why he did this, save to claim that he had done so by mistake because he was “nervous” and “scared” during the Scene Investigation, and that he had been “afraid” to tell ASP Razali that he had made a mistake (see [30] above). But this was not at all consonant with his claim that just a day earlier, his fellow remandees had given him the courage to make alleged “corrections” in his Second Statement (see [28] above).

172 In our view, the patchwork of explanations provided by the Respondent for the inconsistencies in his statements to the police fortifies the District Judge’s findings that the Respondent was trying to tailor his defence to be as close as possible to Nurse MJ’s account: see the DJ’s GD at [63]. Even if the Respondent had managed to identify a possible doubt in the Prosecution’s case, this was not at all a reasonable doubt grounded in the evidence. Moreover, due to the glaring inconsistencies inherent within the Respondent’s defence, the Prosecution had successfully rebutted that doubt.

(3) Nurse MJ’s evidence was externally consistent when juxtaposed against the Respondent’s account

173 For completeness, we note that at [40]–[42] of his Written GD, the Judge alluded to “the inherent probabilities arising from the other factual circumstances” that weighed against Nurse MJ’s evidence. For instance, Nurse DS had walked into the Room about one minute and 40 seconds after Nurse MJ left, and had noticed that the Victim was asleep. This was a “drastic change” from Nurse MJ’s observation that the Victim had been crying in pain.

174 The Judge’s finding may be characterised as an external inconsistency in Nurse MJ’s evidence under the “unusually convincing” standard. However, it bears repeating that the application of this standard must be a holistic one directed at the totality of the circumstances: see [92] above. One such circumstance was the Victim’s mental incapacity, which rendered her unfit to testify in the first place. In this regard, Dr Pamela Ng Mei Yuan (“Dr Ng”) from the Institute of Mental Health opined that as a result of her strokes, the Victim had emotional dysregulation causing mood fluctuations, and would alternate between crying and giggling. On occasion, her disability prevented her from showing emotions even if she was distressed. The Victim was also incapable of explaining what sexual intercourse was, or its consequences. Dr S, the Home’s resident physician, echoed this assessment, stating that the Victim had a poor mental capacity in terms of understanding what was happening to her. The Judge’s comment at [40] of his Written GD that “the drastic change [in the Victim] from crying in pain to being asleep is something that cannot be lightly regarded” was predicated on standard behaviour from the Victim. However, that expectation was neither meaningful nor realistic given the Victim’s cognitive impairments.

175 Another facet of the analysis is that the aforesaid “drastic change” in the Victim should also be assessed against the Defence’s case, which constitutes part of the totality of the circumstances. There was nothing to the point that the Victim had exhibited a rapid change in behaviour, because even on the Respondent’s own case, the Victim was capable of quickly going from being in tears to smiling soon after. This can be seen from the following extract from the Respondent’s oral testimony at the trial:

Q: After you retrieved the round pillow on [the Victim’s] right side, and after you have placed it on [the] left side of her head, how did [she] look?

…

A: I saw that as if tears were coming out of her eyes. It was when I saw her head was touching the side rail and her eyes were – the eyes looked as if tears were coming out, coming out.

Q: Sorry, so just so that I understand. You said that you saw that there were tears coming out of [her] eyes when the side of her head was touching the rail? Did I understand correctly?

A: Yes. Then when I shifted her head towards the centre of the bed, and put the below – pillow beneath her head, then her face was looking normal as if she smiled a little bit.

[emphasis added]

At best, the Victim’s rapid change in behaviour was a neutral factor that ultimately did not assist the Defence in raising a reasonable doubt.

176 In light of the foregoing, we are satisfied that the Judge erred in law, and that his errors had a material impact on his findings. We are further satisfied that the District Judge’s conviction of the Respondent was safe. We therefore exercise our power under s 397(5) of the CPC, which states:

The Court of Appeal, in hearing and determining any questions referred, may make such orders as the High Court might have made as the Court of Appeal considers just for the disposal of the case.

177 In the premises, we reverse the Judge’s order of acquittal and affirm the District Judge’s conviction of the Respondent on the OM Charge.

The sentence for the OM Charge

The framework for outrage of modesty offences under s 354(1) of the Penal Code

178 In sentencing the Respondent to 22 months’ imprisonment (after factoring in seven weeks’ remand) and three strokes of the cane for the OM Charge, the District Judge alluded to the High Court’s decision in Kunasekaran s/o Kalimuthu Somasundara v Public Prosecutor [2018] 4 SLR 580 (“Kunasekaran”), which laid down a two-step sentencing framework for offences of outrage of modesty simpliciter under s 354(1) of the Penal Code.

179 As part of the first step in assessing the offence-specific factors, Chan Seng Onn J laid down the following sentencing bands in Kunasekaran at [49]:

(a) Band 1: less than five months’ imprisonment;

(b) Band 2: five to 15 months’ imprisonment; and

180 In setting out the relevant offence-specific factors, Chan J adopted See Kee Oon J’s approach in the High Court case of GBR v Public Prosecutor and another appeal [2018] 3 SLR 1048 (“GBR”) at [27]–[30], where it was stated that the court should take into account:

(a) the degree of sexual exploitation, which includes which part of the victim’s body the accused person touched, how the victim was touched, and the duration of the outrage of modesty;

(b) the circumstances of the offence, which include considerations of: (i) the presence of premeditation; (ii) the use of force or violence; (iii) the abuse of a position of trust; (iv) the use of deception; (v) the presence of other aggravating acts accompanying the outrage of modesty; and (vi) the exploitation of a vulnerable victim; and

(c) the harm caused to the victim, whether physical or psychological, which would usually be set out in a victim impact statement.

181 After taking into account the offence-specific factors in ascertaining the gravity of the offence, the court should then place the offence within a sentencing band (see GBR at [31]–[38]):

(a) Band 1: This band would include cases with none, or at most one, of the offence-specific factors, and would typically involve a fleeting touch or no skin-to-skin contact, and no intrusion to the victim’s private parts.

(b) Band 2: This band would include cases with two or more offence-specific factors. The lower end of the band would involve intrusion to the victim’s private parts, but no skin-to-skin contact. The higher end of the band would involve skin-to-skin contact with the victim’s private parts, and would also include cases involving the use of deception.

(c) Band 3: This band would include cases with numerous offence-specific factors, especially exploitation of a particularly vulnerable victim, a serious abuse of a position of trust, and/or the use of violence or force on the victim.

182 The court should then consider the offender-specific factors. The aggravating factors might include but are not limited to the number of charges taken into consideration, the accused person’s lack of remorse, and any relevant antecedents demonstrating recalcitrance. The mitigating factors may include a timeous plea of guilt and the presence of a mental disorder or an intellectual disability that relates to the offence (see GBR at [39]).

183 We note that in GBR, See J also set out indicative strokes of the cane for each of the sentencing bands. However, the sentencing framework articulated in GBR concerned outrage of modesty of a minor under 14 years of age, which is punishable under s 354(2) of the Penal Code, whereas the OM Charge in this case is one of outrage of modesty simpliciter under s 354(1). The statutory maximum imprisonment term under s 354(2) is five years’ imprisonment, which is two-and-a-half times that prescribed in s 354(1). The indicative strokes of the cane set out in GBR may therefore not be appropriate in meting out punishments for offences of outrage of modesty simpliciter.

184 We also note that in this regard, Chan J held in Kunasekaran at [50] that the High Court’s decision in Public Prosecutor v Chow Yee Sze [2011] 1 SLR 481 at [9]–[10] “remains instructive in respect of its guidance that the starting point in respect of the imposition of caning is where the outrage of modesty involves the intrusion upon the victim’s private parts or sexual organs” [emphasis in original].

The District Judge’s decision on sentence

185 The District Judge was of the view that the Respondent’s offence fell squarely into the upper end of Band 3 of the Kunasekaran framework due to the multiple offence-specific aggravating factors present in this case. In his view, there was a “high degree of sexual exploitation” [emphasis in original omitted] and abuse of trust. The Respondent had entered a room exclusive to female residents by using his position as a maintenance technician at the Home. Further, he had “exploited the Victim’s extreme vulnerability” [emphasis in original omitted] as he knew that the Victim could not resist his advances due to her severe physical and mental disabilities, and would not be able to raise the alarm by shouting due to her speech impairment: see the DJ’s GD at [95(i)], [95(iii)] and [95(iv)].

186 The District Judge was also of the view that there was significant premeditation on the Respondent’s part. Having helped to set up the audio-visual equipment for the community involvement programme at the Home earlier in the day, the Respondent knew that the remaining residents in the Room were too disabled to raise any alarm even if they were to witness his assault on the Victim: see the DJ’s GD at [95(ii)]. The District Judge also found that “very significant” [emphasis in original omitted] harm had been caused to the Victim, relying on the evidence of Dr Ng, who had interviewed the Victim after the incident. Dr Ng testified that the Victim had shown emotional distress and trauma arising from the incident, and had broken down several times and cried when she was interviewed: see the DJ’s GD at [95(v)].

187 As for the offender-specific factors, the District Judge held that the Respondent had “not shown an iota of remorse”, having made baseless and irrelevant accusations against Nurse MJ and his former supervisor, Mr SBR. In the round, the District Judge considered this “the most shocking and disturbing case” of outrage of modesty he had ever encountered as a district judge, and sentenced the Respondent to close to the statutory maximum term of imprisonment while also meting out three strokes of the cane: see the DJ’s GD at [92] and [96]–[99].

Assessment of the District Judge’s decision on sentence

188 We accept the District Judge’s assessment that there were multiple offence-specific aggravating factors in the present case. In particular, we agree that there was a high degree of sexual intrusion as there was groin-on-groin contact between the Respondent and the Victim. There was a clear abuse of trust as the Respondent was an employee of the Home and had access to the Victim precisely due to his status. We also find it reprehensible that the Respondent had exploited the Victim’s disability, knowing that due to her cognitive and physical impairments, she would not be able to raise the alarm whilst he was sexually assaulting her. On this basis, the Respondent’s offence would fall within Band 3 of the Kunasekaran framework.

189 In our view, the District Judge did not err in his assessment of the offender-specific factors. As the Respondent did not plead guilty, there were no mitigating factors, save for the fact that he had no antecedents. We agree that the Respondent’s spurious accusations against Nurse MJ and Mr SBR, including claims that Mr SBR had committed criminal breach of trust against the Home, were not only irrelevant but also completely unfounded, and showed a lack of remorse on the Respondent’s part. We also agree that given the degree of sexual intrusion, caning was warranted. In this regard, we consider that the sentence of three strokes of the cane was not manifestly excessive.

190 However, we have some concerns about three aspects of the District Judge’s decision on sentence, which we proceed to consider below.

Dr Ng’s psychiatric report

191 First, we note that in his Written GD at [51]–[52] the Judge remarked obiter that it was inappropriate for the District Judge, when sentencing the Respondent, to have referred to Dr Ng’s evidence on the Victim’s emotional trauma and distress arising from the incident, given that the Victim had been found unfit to testify. There should have been caution in referring to the Victim’s reaction and state of mind as to the alleged events.

192 We agree with the Judge’s remarks, though we would caveat that the mere fact that a victim is found mentally unfit to testify does not necessarily mean that evidence of psychiatric or psychological harm to him or her can never be adduced in evidence. The Prosecution can apply for such evidence to be admitted under s 32(1)(j)(i) of the Evidence Act, which states:

32.—(1) Subject to subsections (2) and (3), statements of relevant facts made by a person (whether orally, in a document or otherwise), are themselves relevant facts in the following cases:

…

(j) when the statement is made by a person in respect of whom it is shown —

(i) is dead or unfit because of his bodily or mental condition to attend as a witness …

…

193 Nevertheless, even when such hearsay evidence is ultimately admitted, we would still caution that a court must be minded to guard against any prejudicial effect arising from such evidence, and must test the reliability of such evidence in assigning weight to it.

194 In the present case, the Prosecution did not apply under s 32(1)(j)(i) of the Evidence Act to admit Dr Ng’s psychiatric report. Significantly, in the course of the trial, the District Judge had indicated (and the Prosecution had confirmed) that Dr Ng’s testimony and her psychiatric report would only be adduced for the purposes of assessing the Victim’s fitness to testify, and not for the purposes of assessing the weight to be given to her account of the incident, whether for the purposes of conviction or sentence. Given the District Judge’s indication and the Prosecution’s confirmation during the trial, we are of the view that reliance on Dr Ng’s psychiatric report for the purposes of finding that the Victim had suffered severe psychiatric harm as a result of the incident would be prejudicial to the Respondent, and that the District Judge had erred in taking this report into account.

195 Without any prompting, Mr Kow for the Prosecution quite rightly submitted that it was inappropriate for the District Judge to have relied on Dr Ng’s psychiatric report for the purposes of sentence. He also indicated that the Prosecution would not be applying for the report to be admitted under the hearsay rule. We appreciate the forthrightness of Mr Kow’s concession, which hews to the Prosecution’s role in the fair and impartial administration of criminal justice: see Public Prosecutor v Lim Choon Teck [2015] 5 SLR 1395 at [75].

The evidence of the Respondent’s premeditation

196 Second, we are not fully persuaded that the offence was premeditated by the Respondent, in the sense that he had never intended to enter the Room for the purposes of repairing Mdm JP’s portable television: see the DJ’s GD at [95(ii)]. The District Judge found that Mdm JP’s credit was impeached, given that in her oral testimony in court, she claimed that she had asked the Respondent to repair her portable television on 26 November 2016, the day of the incident, whereas in her statement to the police, she categorically stated that she had not requested the Respondent to do so: see the DJ’s GD at [43]–[46] and [69]–[71].

197 On appeal, the Judge was of the view that the District Judge had erred in impeaching Mdm JP’s credit: see the Written GD at [47]–[49].

198 In our judgment, regardless of the correctness of the impeachment, the overall picture that emerged from Mdm JP’s testimony was that of a very confused individual who was unable to recall whether she had, on the material day, requested the Respondent to repair her portable television. As Mdm JP’s evidence was clearly inconsistent and unreliable, the Prosecution could not positively rely on her statement that she had not made such a request to the Respondent for the purposes of establishing an aggravating factor in relation to sentence. Given Mdm JP’s state of confusion, it was not at all clear whether the true state of events was that documented in her statement to the police, or that narrated in her oral testimony at the trial.

199 We should highlight that this issue makes no difference to the correctness of the District Judge’s conviction of the Respondent. As the District Judge correctly noted, even if the Respondent had truly initially entered the Room to repair Mdm JP’s portable television, he would have had ample opportunity to commit the offence either before or after repairing the television: see the DJ’s GD at [75]. In this regard, we note that the Respondent left the inner section of the Room only at 3.47.07pm, which indicates that he remained at the Victim’s bed for over five minutes after Nurse MJ left.

200 We respectfully disagree with the District Judge that the Respondent premeditated the offence. In our judgment, the most that could be said was that there was a significant degree of opportunism on the Respondent’s part. Whilst he was in the Room repairing Mdm JP’s portable television, the Respondent took advantage of the fact that most of the residents were at the community involvement programme, and that the remaining residents in the Room (such as Mdm MG) were too cognitively impaired to raise the alarm. Such opportunism, whilst an aggravating factor, was not as serious as a finding that the Respondent evinced “considered commitment towards law-breaking” and took “deliberate steps” by scheming to enter the Room for the very purpose of sexually assaulting the Victim: see Ng Kean Meng Terence v Public Prosecutor [2017] 2 SLR 449 at [44(c)].

“[O]ne of the worst cases of its sort”

201 Third, the District Judge considered the present case “one of the worst cases of its sort” and “the most shocking and disturbing case” of outrage of modesty he had ever encountered as a district judge: see the DJ’s GD at [92]. The statutory maximum punishment is reserved for the worst type of cases of the offence concerned: see Public Prosecutor v P Mageswaran and another appeal [2019] 1 SLR 1253 (“P Mageswaran”) at [45] (citing Sim Gek Yong v Public Prosecutor [1995] 1 SLR(R) 185 at [12]). In P Mageswaran at [46], this court noted that a close examination of the nature of the crime and the circumstances of the offender was necessary, and that “[t]o satisfy the criterion of being one of the worst type of cases of [the offence concerned] would generally entail an absence or at least a lack of material mitigating circumstances”.

202 As we noted at [189] above, the present case displayed few, if any, mitigating factors, save for the Respondent’s lack of antecedents. On the other hand, in light of our foregoing analysis, the Prosecution could not establish that there was premeditation on the Respondent’s part, or that the Victim had suffered severe psychiatric harm. This meant that the aggravating factors in relation to the nature of the crime in this case were not as severe as those in other cases: see P Mageswaran at [48].

203 Before the Judge, the Prosecution alluded to the case of Public Prosecutor v ABC [2003] SGCA 14 (“ABC”), which it submitted had similar aggravating factors. In ABC, this court imposed the statutory maximum of two years’ imprisonment on the accused person for a charge of outrage of modesty under s 354 of an earlier edition of the Penal Code (namely, the Penal Code (Cap 224, 1985 Rev Ed)). The accused person, who was the stepfather of the seven-year-old victim, had outraged the victim’s modesty by pushing his penis in and out of her buttocks. He had claimed trial, and it was unclear if he had any antecedents. When viewed in its totality, we are not persuaded that the present case involved offence-specific and offender-specific factors that were as egregious as those that featured in ABC.

204 Of course, in order to warrant the imposition of the maximum sentence, the offence in question does not need to be equally egregious as the offences in the precedents in which the maximum sentence was meted out. What is required is that the conduct of the offender falls within the “range of conduct which characterises the most serious instances of the offence in question” [emphasis in original omitted]: see P Mageswaran at [45]. In our judgment, given the absence of premeditation and psychiatric harm in the present case, this case was not such a case. In our view, the Respondent’s offence fell within the middle of Band 3 of the Kunasekaran framework. After adjusting the sentence upward by taking into account the relevant offence-specific and offender-specific aggravating factors, and after factoring in the seven weeks that the Respondent has previously spent in remand, an appropriate sentence would be 16 months’ imprisonment and three strokes of the cane.

The Court of Appeal’s power to substitute a lower court’s sentence in a criminal reference

205 All that remains is to consider whether, in a criminal reference, this court has the power under s 397(5) of the CPC to reduce the sentence meted out at first instance.

206 In Mok Swee Kok v Public Prosecutor [1994] 3 SLR(R) 134 at [21]–[22], this court considered s 60(4) of the Supreme Court of Judicature Act (Cap 322, 1985 Rev Ed), which was a predecessor provision of s 397(5) of the CPC. This court referred to the UK House of Lords’ decision in Attorney-General for Northern Ireland v Gallagher [1963] AC 349, noting that if its powers under s 60(4) were only confined to answering the question of law posed, there would be no reason to grant it the power to “make such orders as the High Court might have made”. Hence, this court held at [21] that s 60(4) “empower[ed] the court to go beyond answering the questions of law reserved for its determination”.

207 In our judgment, s 397(5) of the CPC grants this court such powers as the High Court in its appellate or revisionary jurisdiction would have had. One such power is the power of the High Court, as an appellate court, to “reduce or enhance the sentence, or alter the nature of the sentence” in an appeal against sentence: see s 390(1)(c) of the CPC.

208 Accordingly, we reverse in part the District Judge’s order as to sentence by substituting the original sentence of 22 months’ imprisonment and three strokes of the cane with a sentence of 16 months’ imprisonment and three strokes of the cane.

Conclusion

209 We are grateful for the assistance of counsel for both parties, and thank them for their detailed submissions. In particular, we appreciate Mr Lau’s pro bono representation of the Respondent under the Criminal Legal Aid Scheme, which he has valiantly and steadfastly maintained from the trial in the District Court, to the appeal to the High Court, to the present Criminal Reference and CM 7 before this court.

Sundaresh Menon Andrew Phang Boon Leong Judith Prakash
Chief Justice Judge of Appeal Judge of Appeal

Kow Keng Siong, Agnes Chan, Chin Jincheng and Etsuko Lim (Attorney-General’s Chambers) for the applicant in Criminal Reference No 6 of 2018 and Criminal Motion No 7 of 2019;
Lau Wen Jin (Dentons Rodyk & Davidson LLP) for
the respondent in Criminal Reference No 6 of 2018
and Criminal Motion No 7 of 2019.

Annex A

SUPREME COURT OF SINGAPORE

22 January 2020

Case summary

Public Prosecutor v GCK and another matter [2020] SGCA 2
Criminal Reference No 6 of 2018 and Criminal Motion No 7 of 2019

-----------------------------------------------------------------------------------------------------

Decision of the Court of Appeal (delivered by Chief Justice Sundaresh Menon):

Outcome: Court of Appeal reverses the respondent’s acquittal and restores his conviction on a charge of outrage of modesty, but substitutes the trial judge’s sentence with a sentence of 16 months’ imprisonment and three strokes of the cane.

Pertinent and significant points of the judgment:

The “unusually convincing” standard is the only applicable standard for both sexual and non-sexual offences where an eyewitness’s uncorroborated testimony forms the sole basis for a conviction. It is incorrect to apply different evidential standards to an eyewitness on the one hand, and an alleged victim on the other: at [87], [98], [99] and [104]).
The evidence of an eyewitness is neither less nor more reliable than that of an alleged victim. No general rule can be laid down about the relative evidential reliability of eyewitnesses compared to victims because this depends on all the circumstances of the case: at [108] and [114].
In assessing the evidence under the “unusually convincing” standard, the true question is how the standard is to be applied, and, more importantly, whether the court has correctly carried out its task of determining whether the Prosecution has discharged its burden of proving its case against the accused person beyond a reasonable doubt: at [125].

Background

1 The respondent (“the Respondent”), a male employee at a nursing home (“the Home”), was charged with one count of outrage of modesty under s 354(1) of the Penal Code (Cap 224, 2008 Rev Ed) (“the OM Charge”). The alleged victim (“the Victim”) was an elderly female resident of the Home who was suffering from serious physical and cognitive disabilities. Because she was found unfit to testify, the Prosecution’s case rested substantially on the testimony of an eyewitness, a female nurse at the Home (“Nurse MJ”). Nurse MJ alleged that she had seen the Respondent straddling the Victim with his trousers pulled down and the Victim’s diaper exposed, and with his groin placed on the Victim’s groin.

2 The Respondent denied the OM Charge. He claimed that he had been asked by one of the residents of the Home (“Mdm JP”) to enter her room (“the Room”), which she shared with the Victim, among other residents, to repair her portable television. While inside the Room, he saw the Victim’s head touching the side railing of her bed and noticed that her pillow was displaced. He thought the Victim was in pain as tears were flowing from her eyes. He therefore placed his knee between the bars of the side railing of the bed to reach for the Victim’s pillow, and then placed it under her head. The Respondent claimed that he did not touch the Victim while he was reaching for her pillow. He testified that after assisting the Victim, he went to the toilet to dispose of some food remains found on Mdm JP’s bed before returning to the Room. A friend of Mdm JP, Ms SBR, came by after he returned to the Room.

The decision at first instance

3 At first instance, the trial judge in the District Court (“the District Judge”) disbelieved the Respondent’s defence and convicted him of the OM Charge. The District Judge found Nurse MJ’s evidence “unusually convincing”, a term that indicates that a witness’s testimony is “so convincing that the Prosecution’s case [is] proven beyond reasonable doubt, solely on the basis of the evidence”. In contrast, the District Judge found the Respondent’s evidence internally inconsistent. Although the Respondent claimed at the trial that he had placed only one of his knees on the Victim’s bed while reaching for her pillow, in one of his statements to the police, he claimed that he had placed both his knees on the bed. During a visit to the scene in the course of the trial, the Respondent was requested to demonstrate how he had retrieved the Victim’s pillow. The District Judge found the Respondent’s posture “highly unnatural and contrived” because the Respondent could easily have reached for the pillow without placing either of his knees on the bed, considering his height.

4 The District Judge considered this case “one of the worst cases of its sort”. After taking into account the seven weeks the Respondent had spent in remand, the District Judge sentenced him to 22 months’ imprisonment (close to the maximum two-year imprisonment term) and three strokes of the cane. In this regard, the District Judge held that there was significant premeditation on the Respondent’s part. He had chosen to strike when several of the Room’s residents were away at an event, with the remaining residents being too disabled to raise any alarm. The District Judge also relied on the evidence of a psychiatrist who had interviewed the Victim after the incident (“Dr Ng”) to find that “very significant” psychiatric harm had been caused to the Victim.

The High Court’s decision

5 On appeal to the High Court, the High Court judge (“the Judge”) acquitted the Respondent of the OM Charge. In so doing, the Judge appeared to suggest the following:

The “unusually convincing” standard applied to an alleged victim’s testimony, but a different standard applied to an eyewitness’s testimony.
The “unusually convincing” standard applied to sexual offences, but a different standard applied to non-sexual offences.
An eyewitness’s observations were of less evidential value than those of an alleged victim.

The Prosecution's criminal reference

6 The Prosecution filed Criminal Reference No 6 of 2018, submitting that the Judge’s decision had given rise to a question of law of public interest. Subsequently, after the Judge issued his written grounds of decision, the Prosecution applied by way of Criminal Motion No 7 of 2019 to reframe the question it had originally posed.

7 The Court of Appeal (“the Court”) agreed to answer the question referred to it, but reformulated it as follows (at [66] and [85]):

What is the standard to be applied when evaluating the evidence of an eyewitness to a crime, where such evidence – (i) is uncorroborated; and (ii) forms the sole basis for a conviction (“the Applicable Standard Question”)?
Is the evidence of an eyewitness inherently less reliable than that of an alleged victim (“the Inherent Reliability Question”)?
How should the court assess the evidence when it applies the “unusually convincing” standard (“the Specific Test Question”)?

The Courts's answers to the questions

The Applicable Standard Question

8 The Court was of the view that the “unusually convincing” standard was the only applicable standard where an eyewitness’s uncorroborated testimony formed the sole basis for a conviction. It was incorrect to apply different evidential standards to an eyewitness on the one hand, and an alleged sexual offence victim on the other. That approach was based on antediluvian views of allegedly “suspect” witnesses such as sexual offence victims and children, and assumed that the evidence of these witnesses was inherently less reliable and therefore had to be treated with more suspicion. The “unusually convincing” standard applied across all types of offences, regardless of whether they were sexual in nature, so long as the evidence presented to the court was uncorroborated (at [87], [98], [99] and [104]).

The Inherent Reliability Question

9 The Court was of the view that the evidence of an eyewitness was neither less nor more reliable than that of an alleged victim. No general rule could be laid down about the relative evidential reliability of eyewitnesses compared to victims because this depended on all the circumstances of the case. The crucial question in each case was whether there was evidence specific to the particular eyewitness or alleged victim concerned which showed that the quality of his or her observations had been compromised or, conversely, improved (at [108], [109] and [114]).

The Specific Test Question

10 As for how the evidence should be assessed under the “unusually convincing” standard, the Court rejected the Prosecution’s submission that the guidelines for identification evidence established in Heng Aik Ren Thomas v Public Prosecutor [1998] 3 SLR(R) 142 (“Thomas Heng”) should apply. The Thomas Heng test was not a prescriptive formula that a judge must invariably recite in deciding whether to convict or acquit, but rather, a set of possible pointers. The true question was how the “unusually convincing” standard was to be applied, and, more importantly, whether the court had correctly carried out its task of determining whether the Prosecution had discharged its burden of proving its case against the accused person beyond a reasonable doubt (at [118], [123] and [125]).

11 The principle of proof beyond a reasonable doubt entailed that upon considering all the evidence presented by the parties, the evidence sufficed to establish beyond a reasonable doubt each and every element of the charge against the accused person. A doubt was not reasonable if it was not reasoned, in the sense of being supported by reasons that were logically connected to the evidence (at [129], [131], [149(b)] and [149(c)]).

12 The principle of proof beyond a reasonable doubt could be conceptualised in two ways:

First, a reasonable doubt could arise from “within the case mounted by the Prosecution”, that is, considering all the evidence adduced by the Prosecution at each stage of the proceedings. The court must particularise the specific weakness in the Prosecution’s own evidence that irrevocably lowered it below the threshold of proof beyond a reasonable doubt. Once the court had identified the flaw internal to the Prosecution’s case, weaknesses in the Defence’s case could not ordinarily shore up what was lacking in the Prosecution’s case to begin with because the Prosecution had simply not been able to discharge its overall legal burden (at [134], [137], [142], [149(e)] and [149(f)]).
Second, a reasonable doubt could arise “on the totality of the evidence”, which included the Defence’s case and any weaknesses therein. The assessment of the totality of the evidence was intimately connected with the “unusually convincing” standard, which arose in situations involving one person’s word against another’s. In order to find a reasonable doubt in the Prosecution’s case on the totality of the evidence, the court had to articulate the specific doubt that had arisen in the Prosecution’s case and ground it with reference to the evidence (at [143], [144], [147], [149(g)], [149(h)] and [149(j)]).

The Court's analysis of the facts

13 The Court held that no reasonable doubt had arisen within the Prosecution’s case, and that the Judge had erred in failing to identify and particularise the flaw within the Prosecution’s case that led him to believe that a reasonable doubt had arisen. The Judge had incorrectly found Nurse MJ’s evidence to be less reliable than the Victim’s evidence. Any weaknesses in a witness’s observations had to be specific to that witness, and there was no suggestion in this case that Nurse MJ’s line of sight had been obscured, or that she had been too far away. On the contrary, she had been standing one-and-a-half arm’s lengths away from the Victim’s bed, and the curtains had been drawn open across the width of the bed. There was also no reason to doubt that Nurse MJ was honest, and nothing inconsistent or implausible about her testimony (at [152]–[159]).

14 The Court also held that no reasonable doubt had arisen on the totality of the evidence. The Respondent had claimed that Nurse MJ was “mistaken” about what she allegedly saw him do to the Victim. But Nurse MJ’s account of the Respondent straddling the Victim on top of her bed was so drastically different from the Respondent’s account of his standing by her bed and assisting her with her pillow that it could not be explained as a mistake. Further, there was a glaring inconsistency in the Respondent’s defence as to the sequence of events. He claimed that after he helped the Victim with her pillow, he went to the toilet to dispose of some food remains found on Mdm JP’s bed before returning to the Room. Ms SBR then came by. The close-circuit television footage showed that those two events occurred at least five minutes before Nurse MJ entered the Room. This meant that on the Respondent’s own account, he would already have reached for the Victim’s pillow and assisted her at least five minutes before Nurse MJ came by, and he therefore could not have been doing the same thing when Nurse MJ saw him. Finally, the Judge had also failed to ground the Respondent’s allegation as to Nurse MJ’s misperception of his posture in the evidence. The Respondent could have easily reached for the Victim’s pillow without placing either of his knees on the Victim’s bed, given his height, and he could not explain the inconsistencies in his statements to the police as to whether he had placed one or both of his knees on the Victim’s bed while he was reaching for the pillow (at [160]–[166] and [168]–[171]).

15 The Court thus reversed the Judge’s order of acquittal and affirmed the District Judge’s conviction of the Respondent on the OM Charge (at [177]).

The Court’s decision on sentence

16 The Court accepted the District Judge’s assessment that there were multiple offence-specific and offender-specific aggravating factors. However, the Court agreed with the Judge that it was inappropriate for the District Judge to have referred to Dr Ng’s evidence when sentencing the Respondent, given that Dr Ng’s evidence had been admitted at the trial only for the purposes of assessing the Victim’s fitness to testify, and not for the purposes of assessing the weight to be given to her account of the incident, whether for the purposes of conviction or sentence (at [191], [192] and [194]).

17 The Court was not fully persuaded that there was premeditation by the Respondent. There was, at most, a significant degree of opportunism, in that he had initially entered the Room to repair Mdm JP’s portable television, but had then taken advantage of the fact that most of the Room’s residents were away at an event and the remaining residents were too cognitively impaired to raise the alarm (at [196] and [200]).

18 The Court explained that the statutory maximum sentence for an offence (in this case, two years’ imprisonment) was reserved for the worst type of cases of the offence. Since the Prosecution could not establish that there was premeditation by the Respondent or that the Victim had suffered severe psychiatric harm, the present case did not fall within that category. After factoring in the relevant offence-specific and offender-specific aggravating factors as well as the seven weeks the Respondent had spent in remand, the Court held that the appropriate sentence was 16 months’ imprisonment and three strokes of the cane (at [201]–[204]).

This summary is provided to assist in the understanding of the Court’s judgment. It is not intended to be a substitute for the reasons of the Court. All numbers in bold font and square brackets refer to the corresponding paragraph numbers in the Court’s judgment.

This judgment text has undergone conversion so that it is mobile and web-friendly. This may have created formatting or alignment issues. Please refer to the PDF copy for a print-friendly version.

Version No 1: 27 Oct 2020 (22:41 hrs)