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
[2019] SGCA 16
Criminal Appeal No 10 of 2018
Between
Public Prosecutor |
… Appellant
And
ASR |
… Respondent
In the matter of Criminal Case No 47 of 2016
Between
Public Prosecutor |
And
ASR |
grounds of decision
[Criminal Procedure and Sentencing] — [Sentencing] — [Intellectually disabled offenders]
[Criminal Procedure and Sentencing] — [Sentencing] — [Young offenders]
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
v
ASR
[2019] SGCA 16
Court of Appeal — Criminal Appeal No 10 of 2018
Sundaresh Menon CJ, Andrew Phang Boon Leong JA, Judith Prakash JA, Tay Yong Kwang JA and Chao Hick Tin SJ
19 September 2018
11 March 2019
Sundaresh Menon CJ (delivering the grounds of decision of the court):
Introduction
1 Writing in 1968, H L A Hart observed of criminal sentencing that “[t]he most difficult problems are presented by young children and the mentally abnormal”: see “Punishment and the Elimination of Responsibility” in Punishment and Responsibility (John Gardner ed) (Oxford University Press, 2nd Ed, 2008) ch 7 at p 184. This appeal was a lucid testament to the truth of this observation as it concerned an offender, the respondent, who was possessed of both characteristics, being just over 14 years of age at the time of the offences which were in issue, and also afflicted with an intellectual disability. Compounding the difficulty of the case was the seriousness of his crimes: he raped a teenage girl in broad daylight and committed serious sexual abuse on her in other ways. On the other hand, but for his age, he might have invoked a statutory defence to criminal liability on the ground of his lack of maturity. Finally, the minimum sentences stipulated in the charges to which he pleaded guilty, the number of those charges, and the existing sentencing regime presented the sentencing court with only two real alternatives which differed starkly: a lengthy term of imprisonment with caning, or reformative training.
2 In these circumstances, the High Court chose the latter sentencing option. After hearing the Prosecution’s appeal for the former to be imposed instead, we agreed with the High Court and dismissed the appeal. In these grounds, we explain our reasons in full as well as discuss the delicate and complex issues of sentencing principle that arose for determination.
Background
3 The respondent was 17 at the time he was sentenced in the court below, and 18 when this appeal was heard. Before he was remanded for the offences in issue, he lived with his mother, grandmother and six siblings in a one-bedroom flat. He was a student at a school for children with special needs, and was assessed by the Institute of Mental Health (“IMH”) a few months after the commission of the offences to have an IQ of 61. His mental age was assessed by one expert to be eight years old, and by another to be between eight and ten years old. It appears that the respondent was just 11 when he first committed an offence. He was thereafter placed on a one-year guidance programme, which he completed. Between the middle of 2013 and 2014, he committed a series of offences which culminated in these proceedings. He was 13 at the time of the earliest of those offences, and 14 at the time of the three sexual offences which were the subject of the charges proceeded with in this case.
The early offences
4 In June 2013, the respondent and three of his friends burgled a flat and stole a number of household items in the total value of $41. He was not charged for this and was instead administered a stern warning in April 2014 on the condition that he not reoffend within the next twelve months. Three months later, in July 2014, he breached that condition by acquiring an EZ-link card of unknown value which he had reason to believe was stolen property. In the same month, he also burgled another flat and stole $300 in cash as well as seven packets of cigarettes. The next day, together with three friends, he stole a mobile phone, six packets of cigarettes and $1,500 in cash from a stranger.
5 The respondent was arrested shortly after. For his conduct in the four incidents mentioned above, he was charged under the Penal Code (Cap 224, 2008 Rev Ed) (“the Penal Code”) with, respectively: (a) theft in a human dwelling with common intention under s 380 read with s 34; (b) dishonest retention of stolen property under s 411(1); (c) house-breaking by night in order to commit theft with common intention under s 457 read with s 34; and (d) snatch theft with common intention under s 356 read with s 34. These were, respectively, the seventh, eighth, sixth and fifth charges in this case.
6 The respondent was then remanded at the Singapore Boys’ Home pending the investigation of these offences. He was released on bail towards the end of July 2014, and proceeded to commit two other offences. In September 2014, he appropriated a friend’s skateboard, and in October 2014, he grabbed the buttocks of a 21-year-old girl. For these actions, he was charged under the Penal Code with criminal breach of trust under s 406 and outrage of modesty under s 354(1) respectively. These were, respectively, the ninth and the tenth charges in this case. He was arrested, and later released on bail again.
The rape
7 On 21 November 2014, the respondent was distributing flyers with his brother and a friend in Bukit Panjang. At about 5.00pm he went to a convenience store to take a break. It was then that he spotted the victim. She was 16 years old at the time, and has been assessed to have an IQ of 50. She was a schoolmate of the respondent, although they did not know each other.
8 The respondent decided to follow the victim because he had become aroused upon seeing her. He tailed her across two pedestrian crossings to the block of flats where she lived. He hid behind a wall while she waited for the lift. When she entered the lift, he hurried in after her. He pressed the button for the highest floor, while she pressed the button for a lower floor. When the doors opened for her floor, she exited into the lobby. He followed her and said, “Baby, I love you.” She did not respond and walked towards her flat.
9 The respondent then pushed her against the parapet. Afraid, she froze. He hugged her and kissed her on her lips and neck. She told him to go away, but he persisted. He then unzipped his shorts and took out his penis. Squatting down, he lifted her dress and pulled her panties to her ankles. He placed his hand inside her bra and felt her breasts. He then inserted a finger into her vagina. She felt pain. He then told her to lie down. She refused and tried to flee, but he restrained her, saying: “If you never lie down now, I take out my knife.” He then pushed her to the floor.
10 Climbing on top of her, the respondent inserted his penis into her vagina. He was not wearing a condom at that time. Again, the victim felt pain. He then ejaculated on her underwear. But her ordeal was not over. He went through her belongings and found a comb that was about 15cm long. He inserted it into her vagina. After taking it out, he placed it into her mouth. He had no reason to believe that she consented to his actions, but decided to have his way because, in his own words, he “felt horny”. He then said, “Bye bye”, and left the scene.
The aftermath
11 The victim returned to her flat and began to cry. Her family later brought her to make a police report. Two days later, the respondent was arrested. His bail was revoked and he has been remanded at the Singapore Boys’ Home since then.
12 For inserting his finger and the comb into the victim’s vagina, the respondent was charged under the Penal Code with two counts of sexual assault by penetration under s 376(2)(a), which is punishable under s 376(3). For threatening to use a knife and penetrating the victim’s vagina with his penis, the respondent was charged under the Penal Code with one count of aggravated rape under s 375(1)(a) read with s 375(3)(a)(ii). These three charges were the third, fourth and second charges respectively in this case. There was originally a first charge for rape, but that was later withdrawn.
13 In April 2015, all ten charges were laid against the respondent in the Youth Court. The first to the fourth charges were transmitted to the High Court. Although the investigations had concluded by then, according to the Prosecution, “the pre-trial process, the need to secure various psychiatric reports, and the need to fix hearing dates that took into account the court[’s] and [the] parties’ availabilities” resulted in the matter being heard only almost two years later in February 2017. By then, the respondent had turned 16 and had spent nearly two years in remand.
The proceedings below
14 Before the High Court judge (“the Judge”), the Prosecution proceeded with the second, third and fourth charges, and, with the respondent’s consent, invited the Judge to take the fifth to the tenth charges into consideration for sentencing purposes. The respondent pleaded guilty to the proceeded charges, and the Judge convicted him accordingly.
The Newton hearing
15 After hearing submissions on the appropriate sentence, the Judge called for a report on the respondent’s suitability for reformative training. Dr Jacob Rajesh, a senior consultant psychiatrist in the Singapore Prisons Service (“the SPS”), prepared and issued a series of memoranda between February and March 2017 in which he maintained that because the respondent had mild mental retardation, he was not suitable for reformative training.
16 The parties then applied in April 2017 for a Newton hearing to determine two issues: (a) the respondent’s prospects of rehabilitation; and (b) his risk of reoffending. At the hearing, the Judge directed the Prosecution to provide statistics on: (a) the historical number of sexual offenders who had been deemed eligible for reformative training, the number who had been sentenced accordingly and the number who had not; (b) for those offenders who had not been sentenced to reformative training, the reasons why; (c) the number of offenders in each of the two categories who had sub-normal IQ; and (d) for those with sub-normal IQ, what had been recommended in their reformative training suitability reports.
17 The statistics showed that the mere fact that a sexual offence was involved did not preclude the imposition of a sentence of reformative training. Out of 830 reformative training suitability reports surveyed, 11 offenders (including the respondent) had been convicted of sexual offences, meaning rape simpliciter, sexual assault by penetration or outrage of modesty. Of these 11, nine had been sentenced to reformative training, one had not and the one remaining (the respondent) had not been sentenced yet. The statistics also showed that the mere fact that an offender had an intellectual disability or a low IQ score did not preclude the imposition of reformative training. Of the 830 reformative training suitability reports surveyed, 19 offenders had sub-normal IQ, that is, borderline IQ of 70–79 and low IQ of below 69. Of these 19, 11 (including the respondent) had been assessed to be unlikely to benefit from reformative training, while eight had been assessed to be likely to benefit from it. Three of these eight had been sentenced to reformative training.
18 After the Newton hearing, the Prosecution pressed for a global sentence of between 15 and 18 years’ imprisonment and at least 15 strokes of the cane. The respondent, on the other hand, submitted that reformative training would be the appropriate sentence in the present circumstances.
The Judge’s decision
19 On 12 March 2018, the Judge sentenced the respondent to reformative training. The Judge’s grounds of decision were published the following month: see Public Prosecutor v ASR [2018] SGHC 94 (“the GD”). In the Judge’s view, the “main issue” before him was whether the respondent should be sentenced to reformative training (at [47]). He answered this in the affirmative because, in his view, rehabilitation was the dominant sentencing consideration, and this made reformative training the more appropriate sentence compared to imprisonment with caning. The GD addressed the main issue through two lines of inquiry: (a) whether the circumstances of the respondent’s offences made reformative training an inappropriate sentence; and (b) whether the respondent’s intellectual disability made that sentencing option inappropriate.
20 The Judge resolved the first question in the negative for the following reasons:
(a) First, the Prosecution was wrong to suggest that the gravity of the respondent’s offences made reformative training inappropriate. The fact that serious offences had been committed had not stopped the courts from applying rehabilitative sentencing options in previous cases (at [59]–[63]). Also, the Prosecution had presented the respondent as being more culpable than he actually was (at [70]). In addition, the statistics showed that the mere fact that an offender had committed sexual offences and had an intellectual disability did not inherently preclude him from being sentenced to reformative training (at [84]).
(b) Second, the respondent’s culpability in relation to the charges that were taken into consideration for sentencing purposes was not as serious as the Prosecution had suggested (at [87]). The Judge found that the evidence showed, among other things, that the respondent’s ninth charge (for criminal breach of trust) was in connection with his failure to return a friend’s skateboard (at [90]), and that he played a “minimal role” in relation to those offences which he was alleged to have committed with others (at [91]–[92]). The Judge also rejected the Prosecution’s view that the fifth and sixth charges showed that the respondent had a propensity to target the vulnerable, reasoning that while the victims in those charges were elderly men, there was no evidence that the respondent had targeted them because of their vulnerability (at [94]–[95]).
(c) Third, the Prosecution was wrong to offer lengthy incarceration as the solution to the significant risk that the respondent would reoffend in the absence of proper rehabilitation, the existence of that risk being a fact which was accepted by the experts on both sides (at [100]–[101]). The Judge doubted that society would be better protected after the respondent was eventually released as he would by then be “stronger and bigger, but lacking insight into the consequences that his choices and conduct carry” (at [101]). Instead, rehabilitation offered a “practical, longer-term solution” which would enhance the protection of the public (at [102]).
21 The Judge also had regard to the trauma that the victim had suffered. But he decided that that had to be balanced against other considerations, including the long-term protection of the public and the rehabilitation of the respondent (at [107]). For all these reasons, the Judge concluded that rehabilitation remained the dominant sentencing consideration (at [108]).
22 Turning to the second line of inquiry, the Judge was not satisfied that the respondent’s intellectual disability made reformative training an inappropriate sentence. The Judge began by setting out his findings on the respondent’s condition. The IMH had found that the respondent functioned in the “extremely low range of intelligence” and had an IQ of 61. This score was equal to or better than just 0.5% of his peers of the same age. His “verbal comprehension” was assessed to be “extremely low”, and his “perceptual reasoning”, “working memory” and “processing speed” were “borderline” (at [112]). His adaptive behaviour skills were in the bottom 1% of young people of his age (at [113]). Based on the foregoing data, the Prosecution’s main psychiatric expert witness, Dr Cai Yiming, an emeritus consultant at the IMH’s Department of Child and Adolescent Psychiatry, accepted that the respondent had a mild intellectual disability (at [114]).
23 Next, the Judge considered the evidence on the respondent’s suitability for reformative training. This was largely evidence in the form of expert opinion. Giving evidence for the Prosecution were Dr Rajesh; Dr Cai; Ms Desiree Choo, a clinical psychologist with the IMH; Mr Ong Pee Eng, a deputy director in the Operations Division of the SPS; Mr Ng Kheng Hong, an assistant director in the same division; and Mr Soh Tee Peng William, a senior assistant director in the SPS’s Correctional Rehabilitation Service Branch. Giving evidence for the respondent were Dr John Bosco Lee, formerly a senior consultant psychiatrist with the SPS and now a private psychiatrist, and Dr Munidasa Winslow, formerly a psychiatrist with the IMH and now a senior consultant psychiatrist in private practice. Having considered their evidence, the Judge concluded that while it would probably be difficult for the respondent to benefit from reformative training because of the way it was usually conducted, adjustments could be made to the programme to meet the respondent’s needs. The Judge summed up the evidence of the main expert witnesses as follows:
(a) Dr Rajesh’s evidence was that the respondent was unsuitable for reformative training because the programme was group-based and catered for inmates with “normal levels of intelligence” (at [119]). Having a low IQ, the respondent would find it difficult to “comprehend” the programme as it was based on changing a person’s cognitive behaviour by formulating reoffending prevention strategies (at [120]). It was not impossible that the respondent would benefit, but the chances of that were “very, very low”, meaning that there was a 75–80% chance that he would not benefit (at [122]).
(b) Dr Lee opined that the respondent should ideally be placed in a mental institution such as the IMH where he could receive special treatment for his condition (at [123]).
(c) Dr Cai opined that the respondent needed to be kept in a structured environment, and that the focus had to be on helping him to overcome his disability in a way that he could follow and understand. Dr Cai doubted whether reformative training would be effective in this respect (at [124]).
(d) Mr Ng’s evidence showed that suitable modifications could be made to the reformative training programme so as to cater to intellectually disabled individuals (at [127]). The Judge noted (at [126]) that in this regard, Mr Ng had referred to the case of Public Prosecutor v Mohammad Fadlee bin Mohammad Faizal [2016] SGDC 274 (“Mohammad Fadlee”). There, reformative training had been imposed on an offender with borderline intelligence who had been convicted of (among other charges) a charge of attempting to insult a woman’s modesty, with the reformative training programme “suitably modified” to suit the offender’s “unique circumstances”.
24 Further, the Judge found that the respondent was not suitable for imprisonment for two reasons. First, the Judge accepted Dr Lee’s opinion that imprisonment could enhance the respondent’s criminality through negative peer influence, and that prison was a stressful environment which might have particularly adverse psychological effects on persons with mental retardation, such as the respondent (at [134]). Second, the Judge rejected the Prosecution’s submission that imprisonment was the preferred option because the respondent would be adequately cared for through a tailored one-to-one rehabilitation programme which the SPS would be able to provide. If a customised programme for the respondent could be provided in prison, the Judge reasoned, it ought to be possible likewise to provide this in the context of reformative training (at [136]).
25 The Judge then observed that apart from reformative training and imprisonment with caning, there were no viable alternative sentencing options. The respondent was not eligible for a mandatory treatment order under s 339(3) of the Criminal Procedure Code (Cap 68, 2012 Rev Ed) (“the CPC”) because he was not suffering from a psychiatric condition (at [142]). Nor was he a person of unsound mind, and therefore, the provisions in Division 5 of Part XIII of the CPC that catered for such persons were not applicable (at [143]). Police supervision after the conclusion of the sentence imposed, under s 309(1) of the CPC, was not an available option because the respondent did not meet the requirement that he have a prior conviction for an offence punishable with imprisonment for two years or more (at [148]).
26 Next, the Judge considered that the respondent had possibly been prejudiced by the nearly two-year delay in the proceedings between the time he was charged in April 2015, when he was 14, and the time of his plea-of-guilt hearing in February 2017, by which time he had turned 16. It appeared to the Judge that if the respondent’s conviction or sentencing had taken place before he turned 16, such that he remained a “juvenile” as defined in s 2(1) of the Children and Young Persons Act (Cap 38, 2001 Rev Ed) (“the CYPA”), he would have been able to seek the exercise of the court’s discretion under s 323 of the CPC read with s 37(2) of the CYPA not to impose a sentence of imprisonment, provided he had not been certified by the court to be “of so unruly a character that he could not be detained in a place of detention or a juvenile rehabilitation centre” (at [159]–[162]). However, the Judge considered that these provisions were unclear as to the applicable date for assessing an offender’s age (which could be, among other possible dates, the date of the commission of the offences concerned or the date of the offender’s conviction), and urged legislative clarification in this regard (at [166] and [168]). In any event, the Judge did not take into account the possible prejudice that the respondent might have suffered. This was because even if there were no such prejudice, the Judge would still have been inclined to impose a sentence of reformative training (at [167]).
27 Finally, the Judge expressed a concern not to undo the progress that the respondent had already made at the Singapore Boys’ Home, where he had been remanded for the preceding three years. The Judge accepted the evidence of the manager of the Youth Guidance Management programme at the Home, Mr Murugasvaran s/o Madasamy, that the respondent was “very well-behaved” and motivated to change, could be trusted, and had shown remorse for his actions (at [172]). This evidence also suggested to the Judge that the respondent was “not as irredeemable as one might have assumed” (at [173]).
28 For these reasons, the Judge sentenced the respondent to reformative training (at [176]). The Prosecution appealed against that decision.
The appeal
29 Before this appeal came on for hearing, we indicated to the parties that we wished to be addressed on whether Parliament intended at all for intellectually disabled offenders who had not attained sufficient maturity of understanding to judge the nature and consequences of their conduct to be liable to be convicted and sentenced to imprisonment, caning or reformative training in the light of s 83 of the Penal Code and any other relevant material. Section 83 provides an absolute defence to criminal liability for any child above seven but under 12 years of age who, at the time of his offending conduct, was not of “sufficient maturity of understanding to judge of the nature and consequence of his conduct”. The parties understood this as an invitation to consider whether s 83 was applicable, and therefore made submissions addressing that issue. We should add here that in the course of the Newton hearing below, the Judge had similarly asked the parties to consider the applicability of this provision, and both parties had taken the position that it was not applicable: see the GD at [153].
30 We also invited submissions on whether the conviction and imposition of a sentence of imprisonment, caning or reformative training on the respondent would in any way violate the Constitution of the Republic of Singapore (1985 Rev Ed, 1999 Reprint) (“the Constitution”). The Prosecution understood this as an invitation to address the constitutionality of the legislative provisions implicated in the respondent’s conviction and sentencing, including the relevant offence-creating provisions, the provisions which created the sentencing options that were available to the court, and the provisions which set out any possible defence that the respondent might have relied upon such as s 83 of the Penal Code. In contrast, the respondent restricted his constitutional challenge to just s 83. We therefore limit ourselves in these grounds to expressing our views on whether s 83 is constitutional.
The Prosecution’s case
31 The Prosecution submitted that s 83 did not apply to the respondent because it applied only to offenders “above 7 years of age and under 12”, and the respondent was 14 at the time of his offences. In particular, that phrase, in the Prosecution’s view, did not encompass the concept of mental age, considering the ordinary meaning and legislative history of s 83, and therefore, the respondent’s mental age of between eight and ten years had no significance in this context. The Prosecution also contended that s 83 was consistent with Arts 9(1) and 12(1) of the Constitution because it employed a differentia, namely, chronological age, which bore a rational relation to the object of the provision. The Prosecution further argued that in any event, the respondent was able to “judge of the nature and consequence of his conduct” within the meaning of s 83.
32 The Prosecution maintained its position that the respondent should be sentenced to a term of between 15 and 18 years’ imprisonment and at least 15 strokes of the cane. Its first principal submission was that the Judge erred by wrongly identifying the relevant and controlling sentencing objectives. It argued that rehabilitation was not the dominant sentencing consideration in this case because the respondent’s offences were serious, the harm caused was severe, the respondent was hardened and recalcitrant, and reformative training was not viable. Instead, retribution and prevention were the key sentencing objectives. The Judge erred in finding that the respondent’s culpability had been affected by his intellectual disability, and failed to appreciate that the various aggravating factors in this case compelled a retributive sentence.
33 The Prosecution’s second principal submission was that the Judge erred in selecting the wrong sentencing option to meet the applicable sentencing objectives. It argued that reformative training was manifestly inappropriate and, indeed, inadequate because it did not meet the need for retribution, and also would not reform the respondent, who was not suitable for reformative training. The Prosecution submitted that imprisonment was the only suitable sentencing option, not only because it would serve the objectives of retribution and prevention, and would be in line with the established sentencing guidelines for rape and sexual assault by penetration, but also because it would enable the respondent to be rehabilitated within a controlled environment and over a longer period of time. An incapacitating sentence was needed, the Prosecution emphasised, to protect the public from the respondent’s high risk of reoffending.
The respondent’s case
34 The respondent contended that s 83 of the Penal Code ought to be read purposively to apply to offenders whose mental age was between seven and 12 regardless of whether their chronological age was above or below 12. If this interpretation of s 83 were accepted, it would mean that the respondent would be entitled to rely on the provision. However, the respondent did not seek to be acquitted because that would result in the withdrawal of rehabilitative possibilities for him. If, on the other hand, the respondent’s interpretation of s 83 were wrong, then s 83, in his view, would be inconsistent with Arts 9(1) and 12(1) of the Constitution because it would be discriminatory for failing to extend the protection of the statutory defence to offenders who, although not of the requisite chronological age, might nonetheless be of insufficient maturity of understanding as a result of their mental age.
35 The respondent adopted the reasons of the Judge in seeking to uphold his decision to impose reformative training. The respondent presented two principal sets of arguments. The first was a collection of general principles, including the propositions that the intellectually disabled should not be precluded from benefiting from reformative training, that the assessment of an offender’s culpability should be affected by his IQ, and that the seeming gravity of an offence should not automatically preclude the imposition of reformative training. The second set of arguments sought to highlight how the specific features of this case justified the Judge’s sentence of reformative training. In this regard, the respondent argued that the sentence was consistent with precedent, that his culpability as reflected in the charges taken into consideration did not demonstrate escalating criminal behaviour, and that reformative training would be in both his and society’s long-term interests.
The issues on appeal and our approach
36 In that light, we outline the five broad issues which, in our judgment, arose for determination in this appeal.
37 The first was the extent of the respondent’s intellectual disability. This appeal in large measure turned on the proper characterisation of the respondent’s mental state at the time of his offences. It was therefore appropriate to begin with a close examination of this issue in order to establish a proper view of the factual matrix of this case.
38 The second broad issue was the applicability of s 83 of the Penal Code. Section 83 provides a complete defence to criminal liability if the offender is “a child above 7 years of age and under 12” and “has not attained sufficient maturity of understanding to judge of the nature and consequence of his conduct” on the occasion of his offence. The respondent contended that s 83 applied to him, so we had to consider whether that was correct. As we shall explain, we disagreed with the respondent. On this hypothesis, the respondent submitted that s 83 must then be unconstitutional, but we also did not accept that argument, as we shall explain.
39 The third broad issue was the appropriate sentencing framework for intellectually disabled young offenders convicted of serious offences. This was an important issue because it concerned the general framework for assessing the whole case. It invited the question whether the two-step framework for sentencing young offenders articulated in Public Prosecutor v Mohammad Al-Ansari bin Basri [2008] 1 SLR(R) 449 (“Al-Ansari”) and affirmed in cases such as Public Prosecutor v Koh Wen Jie Boaz [2016] 1 SLR 334 (“Boaz Koh”) was suitable for the sentencing of intellectually disabled young offenders who have committed serious offences.
40 The fourth broad issue was whether rehabilitation should remain the dominant sentencing objective in this case. This required us to examine whether and to what extent deterrence, incapacitation and retribution applied to the respondent in view of his youth and the extent of his intellectual disability.
41 The fifth and final issue was whether, in the light of the applicable sentencing objectives, reformative training was the appropriate sentencing option in this case. This required us to analyse the suitability of, respectively, reformative training and a long imprisonment term with caning, the latter being the Prosecution’s sentencing position.
42 We now turn to address these issues in sequence.
Issue 1: The extent of the respondent’s intellectual disability
43 The extent of the respondent’s intellectual disability centred around three factual issues: first, the significance of his IQ score; second, the significance of his mental age, which entailed consideration of the meaningfulness and reliability of the concept of mental age for sentencing purposes generally; and third, the degree to which his intellectual disability affected his ability to control his impulses. Below, we address each issue in turn, and explain why, together, they demonstrated that the respondent’s intellectual functioning was seriously impaired.
The respondent’s IQ
44 The respondent’s IQ test was administered when he was about 14 and a half years old by Ms Choo. The version of the IQ test which she administered was the fourth edition of the Wechsler Intelligence Scale for Children (“the WISC”), known as “the WISC-IV”. The WISC-IV, according to Ms Choo’s report, is “used to test the general thinking and reasoning skills of pupils aged six through sixteen” and “provides an overall measure of intellectual ability” [emphasis in original omitted].
45 The WISC-IV provides four “index scores”, each of which measures a different aspect of the candidate’s cognitive ability. The four indices are the Verbal Comprehension Index, the Perceptual Reasoning Index, the Working Memory Index and the Processing Speed Index, which are abbreviated in Ms Choo’s report as “VCI”, “PRI”, “WMI” and “PSI” respectively. According to Ms Choo, VCI measures the ability to understand verbal information, to think and reason with words, as well as to express thoughts in words; PRI measures perceptual organisation, which includes the ability to think in visual images, to manipulate such images, to reason without the use of words, and to perceive and interpret visual material; WMI measures the ability to maintain information in conscious awareness, to perform some operation or manipulation of it, and to produce a result; and PSI measures the speed of mental and graphomotor (meaning handwriting) processing, visual-perceptual discrimination, and eye-hand coordination.
46 The candidate’s four index scores are averaged to yield what is called his Full Scale IQ or “FSIQ” score. It is this overall metric that we have in mind, as did the Judge, when we refer to the respondent’s IQ score. The respondent’s WISC-IV test results, including his individual index scores and his overall FSIQ score, were set out in Ms Choo’s report in the form of the following table:
47 In her report, Ms Choo stated that the respondent’s FSIQ score should be “interpreted with caution”, and that his intellectual functioning would be “more meaningfully understood by his performance on the separate indices” because there was “a significant difference between his scores on the individual subtests within this domain”. Explaining the significance of each of the respondent’s index scores, she stated that he scored equal to or better than 0.1%, 6%, 3% and 5% of his same-aged peers in VCI, PRI, WMI and PSI respectively. In its written submissions, the Prosecution echoed Ms Choo’s caution in relation to the discrepancy between the respondent’s VCI score and his other index scores, and suggested that the respondent might be more intelligent than his FSIQ score conveyed.
48 In our judgment, the caution sounded by Ms Choo did not assist the Prosecution because it did not tell us precisely how the discrepancy between the respondent’s VCI score and his other index scores shed light on his state of mind at the time he committed his offences. In particular, although it was clear that each of the four index scores measured a different aspect of the respondent’s cognitive ability, it was not demonstrated to us how each aspect related to the other aspects, and how any of those relationships explained the respondent’s state of mind at the time of his offences. In fact, Ms Choo herself was not prepared to restrict the respondent’s low cognitive functioning to just his verbal comprehension ability. Thus, during cross-examination, she was unwilling to confirm that despite the respondent’s “extremely low” VCI score, his “borderline” PRI score meant that he could think and solve problems without the use of words:
Q Okay. Also, going back to PRI, perceptual reasoning, him scoring borderline, would you say that that shows on some level he is able to think and solve problems, just that the manner in which he does it is through [the] use of images, pictures and not so much words?
A Think that might be al---a bit of a leap from the results and I’m not able to confirm that based on these results.
49 In the circumstances, we concluded that there was nothing improper in reasoning, as the Judge did (see the GD at [112]), on the footing that at the time of the offences, the respondent: (a) functioned in the “extremely low range of intelligence”, according to Ms Choo’s assessment of his overall IQ score of 61; and (b) scored equal to or better than just 0.5% of his same-aged peers.
The respondent’s mental age
50 Next, we consider the respondent’s mental age. The formula for calculating mental age, as stated in a report prepared by Dr Cai, is: “Mental age = (IQ x Physical Age) ÷ 100 (average IQ score)”. Based on his FSIQ score of 61, the respondent, according to Dr Lee, had a mental age of between eight and ten at the time of his offences, and, according to Dr Cai, had a mental age of eight at that time. In the court below, the Judge did not appear to base any part of his reasoning on the respondent’s mental age. He recorded Dr Cai’s criticism that mental age was “archaic and not helpful and meaningful in clinical practice and forensic evaluation”, but neither accepted nor rejected this opinion: see the GD at [150]–[151]. Nonetheless, part of the Prosecution’s case on appeal was that the concept of mental age was not reliable for the purpose of sentencing. For the reasons that follow, we considered the Prosecution’s criticism to be overstated, and found that the concept was in fact a useful heuristic tool for that purpose, although it must be understood in the context of the offender’s life experiences.
The concept of mental age
51 The Prosecution submitted, relying on some scientific literature, that the mental age of a person did not indicate his social and emotional functioning, and did not account for his life experiences. This point was indirectly supported by Dr Cai’s report, which suggested that it was inaccurate to think of the respondent as an eight-year-old boy because such a boy would typically be unable to conceive of or carry out a rape offence. The Prosecution also highlighted that the respondent’s expert, Dr Lee, accepted that even people who were of the same chronological age could exhibit different levels of maturity due to their living or family circumstances. The Prosecution referred to the opinion of one Dr Robert Adler, cited with approval by the Supreme Court of Tasmania in Dobson v Jackson [2009] TASSC 118 at [255] per Evans J, that mental age simply meant that persons with an intellectual disability, for example, with an IQ of 50, “perform in some areas like a nine year old”.
52 We had three observations on these submissions. First, mental age is predicated on the concept of FSIQ. Therefore, it must, as a matter of logic, carry any limitation that FSIQ carries. As we have seen, FSIQ measures verbal comprehension, perceptual reasoning, working memory and processing speed: see [45]–[46] above. FSIQ does not measure social or emotional functioning; nor does it take into account a person’s life experiences, at least not directly. Since mental age is predicated on FSIQ, it cannot be criticised for not taking these factors into account because it does not purport to be an estimate of these factors. What mental age purports to be is an estimate of what a person of a particular chronological age would usually be capable of achieving in terms of verbal comprehension, perceptual reasoning, working memory, and processing speed. Since this is all that the concept of mental age purports to estimate, it cannot then fairly be criticised for failing to be an estimate of more than that.
53 Second, as mental age does not take into account a person’s life experiences, it follows that the older a person, the less probative value his mental age would have, all other things being equal. Thus, a 40-year-old man who is clinically assessed to have a mental age of an eight-year-old cannot really be said to be thinking and operating like an eight-year-old. This is because he will have lived some 20 or so years as an adult, and may perhaps even have held down a job – all these experiences cannot be discounted when assessing his cognitive ability. Conversely, by the same token, the younger a person, the more probative value his mental age would have, all other things being equal. Thus, take a child who is 12 at the time of his offences, and who is assessed at that time to have a mental age of between eight and ten years old: in such a case, it seems unlikely that the child’s life experiences would have overtaken his mental age to such an extent that his mental age has to be substantially discounted or even disregarded altogether. His situation is quite different to that of a 40-year-old man with a mental age of eight.
54 Third, as a matter of common sense, it cannot be wrong to analogise a person’s maturity or cognitive ability with that of someone who is of a different chronological age. Based on a person’s words and conduct, it is entirely reasonable to say that he is “childish” or that he is “wise beyond his years”. Indeed, implicit in Dr Cai’s opinion that rape is something that “[a]n 8 years [sic] old boy, typically, is not able to think of or carry out” is a notion of what a typical eight-year-old is capable of. The concept of mental age, in our judgment, is simply a heuristic tool for quantifying that kind of analogy in more precise terms. In so far as it is predicated on FSIQ, it carries the limitations associated with FSIQ, and must be employed with an awareness of those limitations. However, it should not be rejected altogether as an unreliable and problematic concept.
55 Nor is any such rejection evident in the scientific literature. Dr Cai explained the origin of the concept of mental age in these terms in his report:
Mental Age is a concept related to intelligence. It was first defined by the French psychologist Alfred Binet, who introduced the intelligence test in 1905, with the assistance of Theodore Simon. Historically, its use was for educational streaming of the French schoolchildren.
It looks at how a specific child performs intellectually, compared to average intellectual performance for that physical age, measured in years.
56 Consistent with this explanation, modern psychological literature continues to employ the concept of mental age precisely to approximate a child’s intellectual performance to that of his similar-aged peers. Thus, for example, in Frank D Baughman et al, “Common mechanisms in intelligence and development: A study of ability profiles in mental age-matched primary school children” (2016) 56 Intelligence 99, researchers examined the relationship between individual differences and cognitive development by comparing the cognitive profiles of groups of younger and older children matched on overall mental age using standard tests of intelligence, including the third edition of the WISC. Notably, the conclusion was that there was a high degree of similarity between these two age groups in terms of cognitive profile, suggesting that children of different chronological ages who are of the same overall ability level in terms of mental age are at a similar developmental and intellectual level.
57 Therefore, contrary to the Prosecution’s submission, we considered that the concept of mental age is far from heterodox, outdated or unreliable. It is certainly true that it must be understood in the context of an offender’s life experiences and circumstances for the purposes of assessing his culpability with regard to the specific offending act(s) that he is said to have committed. But that is simply a function of what the concept of mental age itself purports to represent, and does not in any way reduce its inherent credibility or potential utility as a psychological concept.
Cases on mental age
58 In addition, the cases support the foregoing analysis of the concept of mental age. Indeed, as far as we know, none of the cases, not even the ones cited by the Prosecution, contain an outright rejection of the usefulness of this concept in the context of sentencing. Those cases which do not place significant reliance on it are cases in which the history of the offender’s life experiences shows that he cannot be treated as if he were a child even though he has the mental age of a child. We outline two of these cases below.
59 In Regina v Gordon Laxton [2010] EWCA Crim 2538 (“Laxton”), which was cited by the Prosecution, the offender suffered brain damage at birth and was “markedly impaired in his mental development” (at [4]). At the age of 52, he pleaded guilty to multiple counts of indecent sexual assault perpetrated over a number of years against his cousin, who was then aged between seven and 11, as well as against his niece, who was then aged between seven and 12. These offences were committed approximately from the time when the offender was 22 until he was 36. The first instance judge imposed a community order with a supervision requirement for three years. The Attorney-General appealed, arguing that the sentence was unduly lenient. The English Court of Appeal agreed and substituted a sentence of three years’ imprisonment, holding that although the judge was right to pay close attention to the offender’s “difficulties and disabilities”, he failed to consider the seriousness of the offences (at [19]). The court noted that the offender had been assessed to have a mental age of eight, but disregarded this on the basis that his ability to obtain gainful employment for over thirty years suggested that he was more mature than that. Lord Judge LCJ made these remarks at [13]:
Full investigations have been made as to the extent of the offender’s mental problems. A forensic psychologist was asked by those acting for the offender to assess his intelligence and his suggestibility. The result of verbal and non-verbal intelligence tests produced a score which would have put the offender in the bottom one per cent of the population. The conclusion expressed by the psychologist is that the offender
“is an extremely limited individual who probably has a mental age of approximately 8 years. He would be unable to function as an adolescent, never mind as an adult.”
We approach that conclusion with serious reservation because an 8 year old would not be able to be in gainful employment, even of a modest kind, for thirty years.
60 The next case is Regina v Myles Williams [2013] EWCA Crim 933 (“Williams”), where the offender had an IQ below 70. He was convicted of the murder of a girl, and of wounding with intent two other victims on the same occasion. He committed these offences when he was 19. The first instance judge sentenced him to life imprisonment for the murder, with a minimum term of 28 years’ imprisonment, and also to a concurrent term of 15 years’ detention in a young offenders’ institution for the wounding offences. The offender appealed to reduce the minimum 28-year imprisonment term, relying on, among other things, his mental age, which was assessed as “early teens” (at [10]), and his immaturity. Dismissing the appeal, the English Court of Appeal held that although some allowance had to be given to the offender’s mental age, that was not decisive because, by his conduct, he had shown himself to be capable of a higher level of functioning. In this regard, the court noted that he “liked partying, casual sex and taking illegal substances”, “had fathered children with three women and had held down a job for some two years” (at [14]). Treacy LJ said (likewise at [14]):
As stated, this appellant was 19 at the time of these crimes. The psychological evidence showed that he was of limited intelligence. However, he was able to understand the difference between right and wrong. Although he had a mental age in the early teens, that was only one way of assessing his social functioning. There was evidence to show that he functioned at a higher level than that. The psychologist’s tests were primarily directed to cognitive ability. This appellant, the evidence showed, liked partying, casual sex and taking illegal substances. He had fathered children with three women and had held down a job for some two years. In addition, the judge cited an example of his cunning in the way he had shaped an explanation to deal with inconvenient evidence. The conduct of which the appellant was guilty was gross and in our judgment the judge was correct in his sentencing remarks to disavow any correlation between the level of IQ and what was done in this case. Nonetheless, allowance was to be made for the appellant’s age, which of itself implies a level of immaturity, coupled with his intellectual difficulties which undoubtedly reduced his level of mature social functioning, although not to the point indicated by the IQ tests themselves.
61 Laxton and Williams may be contrasted with R v MBQ, ex parte Attorney-General of Queensland [2012] QCA 202 (“MBQ”), a case cited by the respondent. The offender there was an Aboriginal boy who was 12 when he raped as well as had unlawful and indecent dealings with a three-year-old toddler, and 14 when he was sentenced. He was assessed to be functioning in the range of a nine-year-old at the time of his offences. He pleaded guilty to the offences, and the first instance judge sentenced him to three years’ youth probation with the special condition that he attend a rehabilitative programme as directed by the State. The Attorney-General appealed, contending that the judge had placed excessive weight on the offender’s mental age and other mitigating factors, and had failed to recognise the seriousness of the offences. It was submitted that imprisonment for a term of between three and five years would be appropriate. Dismissing the appeal, the Queensland Court of Appeal accepted that while the sentence seemed lenient, it was justified by the offender’s reduced culpability in the light of his mental age at the time of his offences. Margaret McMurdo P had regard to the evidence of the offender’s intellectual capacity based on psychiatric reports (at [10] and [14]–[15]), and concluded that it indicated that he had a “limited grasp of the consequences and moral blameworthiness of his actions at the time he committed the offences”, and this was “highly relevant” to the sentencing exercise (at [44]).
62 The difference between Laxton and Williams on the one hand and MBQ on the other seems to us to lie in this: in the former two cases, the offenders had accumulated various mature experiences in life which made it difficult for the court to take at face value the assessment that they had the mental age of a child. In contrast, in MBQ, there was no evidence of such experiences. The offender was 12 when he committed the offences and 14 when he was sentenced, so he would not have had the time to experience anything which might suggest that his mental age of nine at the time of his offences was a mischaracterisation of his maturity. Instead, the evidence showed rather the opposite, in that there appeared to be environmental reasons why the offender’s mind had not properly developed, intellectually and socially.
63 In our judgment, therefore, the Prosecution was mistaken in arguing that the court in Laxton “rejected the concept of mental age”. A person’s mental age is, as we have noted, not a complete proxy for his cognitive ability, and what the courts in Laxton and Williams found was simply that evidence of what the offender had shown himself to be capable of doing meant that the court could not take his clinically-assessed mental age to represent that in terms of his cognitive ability, he was like a child. Where, however, there is little or no evidence of that sort, then, in our judgment, the court would ordinarily be justified in relying on an offender’s mental age as a heuristic tool for assessing his degree of culpability.
The significance of the respondent’s mental age
64 In this case, the evidence was that based on his FSIQ score of 61, the respondent had a mental age of between eight and ten at the time of his offences. The significance of this was put in the following terms by Dr Lee during the trial:
Witness: … People with mild mental retardation often have IQ that is measured between 60 to 69. It is important for us to realise that, actually, there [is] quite a substantial population among us because mental retardation prevalence in [the] population is between 1 to 2 percent and mild mental retardation is actually the greatest proportion of this group of people.
A good way to understand the level of functioning of people with mental retardation is to picture someone between 8 to 10 years old, that kind of intellectual. So you are looking at a Primary 2 to a Primary 4 person, that kid [sic] of level of understanding of abstract concepts of judgements. That is essentially the level of cognitive functioning and this is essentially a guide that we usually use, about 9---8 to 10 years old. Mild mental retardation.
Court: No, I don’t follow you. When you say “8 to 10”, what do you mean? You look at children of that age or you compare the patient with children of that age?
Witness: That means that the cognitive---the level of understanding, their level of judgement is comparable to a person who is about 9 years old. A child who is about 9 years old. So, essentially, you know, “Oh, do you understand the concept of peace for a 9-year old? You know, could you---do you understand the concept of police will catch you, 9-year old? Do you understand the police of---cannot steal things?” You know, a 9-year old kind of understanding, not somebody who is 16 years old, somebody who is 20 years old. So it’s, like, comparable to that level and this will give us a very good understanding of what kind of functioning we are looking at for a men---mild mental retardation person. Just like 9-year olds, they can go and buy their own things, they can take the transport. So it’s about that level, Your Honour.
65 For the reasons given at [52]–[63] above, we preferred Dr Lee’s testimony to Dr Cai’s wholesale rejection of the concept of mental age, and would apply mental age as a heuristic tool to evaluate the respondent’s culpability, bearing in mind its limitations.
66 On that analysis, we considered that there was no evidence to indicate that the respondent’s mental age should not be taken to be a reasonably accurate approximation of his cognitive ability at the time of his offences. The respondent was only 14 at that time. There is no evidence that he had by then accumulated life experiences and participated in activities which would suggest that he was in fact more mature than his mental age indicated. We noted that the respondent’s prior offending conduct might suggest that, to some degree, he was no babe in the woods. But we did not find this significant. Several of the offences which the respondent committed before the aggravated rape and the sexual assault by penetration offences were relatively minor. Also, consistent with the Judge’s findings (see the GD at [91]–[92]), the objective evidence suggested that the respondent had simply participated in those offences under the bad influence of his friends. Indeed, the two reformative training suitability reports on the respondent identified his companions as a risk domain in these terms:
Companions: This domain presented with risk.
[The respondent’s] association with his antisocial peers appeared to have increased his risk of offending. He reported close association with his neighbourhood friends, some of who [sic] were aged above 30, who had criminal antecedents and were his accomplices in his past offences. His narrative highlighted that some of these peers also had sexual offending antecedents of outrage of modesty and underaged sex. Furthermore, he reported learning about sexual acts from the sharing of their sexual offences and other sexual experiences.
67 In all the circumstances, we did not think that the respondent’s mental age should be disregarded in the way that the mental age of the offenders in Laxton and Williams was given little weight owing to their life experiences (including their having held down a job) which showed that they were in fact more mature than their mental age indicated. The Prosecution in this case referred to the fact that the respondent was capable of playing with his friends and helping his mother to distribute flyers, but that is of little significance because these are things which one might readily expect an average child of between eight and ten years of age to be capable of doing. We therefore concluded that it was meaningful to consider that the respondent had a mental age of between eight and ten at the time of his offences, and we explain below the effect of this conclusion on our assessment of his culpability.
The respondent’s ability to control his impulses
68 We turn now to the third and last component of the extent of the respondent’s intellectual disability, namely, his ability to control his impulses. The Prosecution accepted that the respondent’s ability to control his impulses had been impaired by his intellectual disability. There was therefore no controversy over this factual point. The dispute pertained, rather, to the proper conclusion to be drawn from such impairment for sentencing purposes, meaning whether the respondent ought to be incapacitated through a lengthy term of incarceration because he is, in the Prosecution’s words, “a sexual predator who is unable to control his urges”, or whether he ought to be rehabilitated because, in the end, he remains a young offender who has demonstrated some progress while in remand. We shall explain our views on this issue later in these grounds. For now, we consider it useful to examine the evidence of the respondent’s lack of impulse control in order properly to establish it as a fact, given that it was relevant to our assessment of his culpability, and because the Judge made only a passing reference to it: see the GD at [99].
69 The principal piece of evidence of the respondent’s lack of impulse control was the opinion of Dr Winslow and his colleague, Dr Amita Sarkar, who conducted a psychological assessment of the respondent and prepared a corresponding report dated 13 July 2015. The respondent was 14 at the time of the assessment. In their report, Dr Winslow and Dr Sarkar opined that while the respondent knew that what he had done to the victim was wrong, his low IQ and poor adaptive behaviour likely contributed to his impulsive behaviour. They put it in these terms:
22 If he suffers from any psychiatric condition or any abnormality of mind;
[The respondent] is a young person with limited cognitive and intellectual abilities. [The respondent’s] IQ & cognitive assessment report as well as his adaptive behaviours assessment confirm that he is functioning much lower than his chronological age. In addition he presents with indications of executive brain function deficits. Executive functions are the skills that an individual of any age must master to deal with everyday life. Self-monitoring is particularly important in later childhood and adolescence, because it governs a person’s ability to evaluate his own behaviour in real time. Individuals with extremely low cognitive abilities and executive functions deficits struggle to think ahead, problem solve or plan their actions. They often act without thinking of possible consequences of their actions.
23 If so, is there a causal link between his psychiatric condition(s) and the offences committed or if these psychiatric conditions are contributory factors;
[The respondent’s] low IQ & cognitive functioning as well as evidence of executive functions deficits are likely to be contributing to his difficulties of response inhibition and impulsive actions, such as alleged offence. He seems unable to think of consequences of his behaviours and hence acts without thinking.
[original emphasis in bold; emphasis added in italics]
70 During the Newton hearing, the Prosecution invited Dr Cai to comment on this part of Dr Winslow’s and Dr Sarkar’s report. Dr Cai agreed that the respondent’s low IQ was a contributing factor in relation to his inability properly to control his impulses to commit criminal acts:
Q … Let’s look at paragraph 23.
...
Dr Winslow opined that:
[Reads] “[The respondent’s] low IQ & cognitive functioning as well as evidence of executive functions deficits are likely to be contributing to his difficulties of response inhibition and impulsive actions, such as alleged offence. He seems unable to think of consequences of his behaviours and hence acts without thinking.”
Do you have any comment on Dr Winslow’s conclusion over here in paragraph 23?
A Here, he implies a low IQ compromise [the respondent’s] impulse in the committing of offence. But I would think this should not overlook the background of conduct disorder in him that make him commit the offence. Suppose he’s just a low IQ without a criminal thinking or behaviour or thought, he might not have committed the offence. But the low IQ helps to lower the---the impulse control and the social judgement in proceeding with the offence. So it is implied he may not seems [sic] to think of---think of the consequences of his behaviours and the acts without thinking. But he was capable in [sic] knowing what he did at the time of the offence.
Q Yes. So to answer the ques---
Khoo: Sorry, just---Your Honour, just one last question before I move on, before we apply for stand down, Your Honour.
Q Dr Cai, to answer this---can you answer this question? Is there between his intellectual disability and his---the commission of the offences? Is there a causal link? What’s the link? Is there any link between his low ID [sic]---
A As I said, the link is because of his conduct disorder, the propensity to commit offence. And this low IQ would add on to lower the judgement impulse control in the commission of offence. In this sense, together, it would make him less able to control from committing offences.
Q I see. Yes.
71 The fact that Dr Cai accepted that the respondent’s low IQ affected his ability to control his impulses is significant. This is because it is precisely this kind of causal link between an impairment of the mind and the commission of criminal offences that the courts have consistently recognised as attracting mitigating weight: see Public Prosecutor v BDB [2018] 1 SLR 127 at [72]; Public Prosecutor v Chia Kee Chen and another appeal [2018] 2 SLR 249 at [112]; Chan Chun Hong v Public Prosecutor [2016] 3 SLR 465 at [73]. An illustration of this principle at work is the High Court’s decision in Public Prosecutor v Chong Hou En [2015] 3 SLR 222, which concerned the question whether voyeurism was a psychiatric disorder, such that the accused’s sentence of probation was justified for the purposes of promoting his rehabilitation, or whether he should have been given a sentence of imprisonment to deter him from reoffending. In explaining the type of disorder that voyeurism had to be proved to be in order to be of mitigating value, Chan Seng Onn J drew on the jurisprudence on sentencing theft offenders who suffered from kleptomania, and held that voyeurism had to be an impulse control disorder in order for it meaningfully to reduce the accused’s culpability. Chan J put it in this way (at [33]):
… [I]t is clear that both [Public Prosecutor v Goh Yee Lin [2008] 1 SLR(R) 824 (“Goh Yee Lin”)] (explicitly) and [Lim Hock Hin Kelvin v Public Prosecutor [1998] 1 SLR(R) 37 (“Kelvin Lim”)] (implicitly) had examined the nature of kleptomania and paedophilia respectively before deciding on the mitigating value to be attached to the disorder. Both kleptomania and paedophilia manifest themselves in the very act criminalised. The High Court in Goh Yee Lin and the Court of Appeal in Kelvin Lim came to different conclusions on the nature of kleptomania and paedophilia respectively and thus ascribed differing mitigating values to each of the mental disorders. The High Court attached significant mitigating value to the diagnosis of kleptomania because of the “undeterrability” of the disorder given that it is an impulse control disorder and the “sufferer may not be fully able to control his or her actions prior to and while committing the offence”. In the light of this, deterrence was rendered less effective and rehabilitation formed the primary focus. The Court of Appeal, on the other hand, attached little or no weight to the disorder of paedophilia since it rejected any “suggestion that the sufferer cannot help it and therefore carries only a diminished responsibility for his actions”. The Court of Appeal further concluded that there was “no evidence that paedophiles cannot exercise a high degree of responsibility and self-control”. Both the High Court and the Court of Appeal were focused on the self-control or lack thereof of the sufferer of the disorder. This to my mind is also the correct inquiry. … [original emphasis in bold italics; emphasis added in italics]
72 In the present case, the report prepared by Dr Winslow and Dr Sarkar and Dr Cai’s oral testimony constitute objective evidence that the respondent’s self-control was impaired by his intellectual disability. While the Prosecution accepted this, it somewhat surprisingly also submitted that the respondent’s intellectual disability “was not causally linked to the commission of his offences”. We rejected this submission because it was contradicted by the evidence and by the Prosecution’s own concession. In our judgment, there was a causal link between the respondent’s low IQ and his commission of the offences, and this attenuated his culpability, as we shall elaborate at [107] below.
Issue 2: Section 83 of the Penal Code
73 The next broad issue is s 83 of the Penal Code. This provision reads:
Act of a child above 7 and under 12 years of age, who has not sufficient maturity of understanding
83. Nothing is an offence which is done by a child above 7 years of age and under 12, who has not attained sufficient maturity of understanding to judge of the nature and consequence of his conduct on that occasion.
74 As we have mentioned, although the Judge invited the parties to address the applicability of s 83 in the course of the Newton hearing below, they declined to do so, taking the view that the provision did not apply: see [29] above and the GD at [153]. Revisiting the issue now in response to our queries, the Prosecution maintained its position below, while the respondent argued that the word “age” in s 83 ought to include mental age, and that s 83 therefore applied to him as he had a mental age of between eight and ten at the time of his offences. We agreed with the Prosecution’s interpretation of s 83, and we also agreed with the Prosecution that, so construed, s 83 was not unconstitutional. That said, we also thought that s 83, when considered together with its neighbouring provisions, shed some light on how a young offender’s culpability should be analysed as a general matter.
The meaning of “age” in s 83
75 The respondent contended that the word “age” in s 83 should be construed to include mental age, such that an offender with a mental age of between seven and 12 would be eligible to rely on the provision even if his chronological age was above 12. The respondent developed this argument within the three-step framework for purposive interpretation set out in Attorney-General v Ting Choon Meng and another appeal [2017] 1 SLR 373 (“Ting Choon Meng”) at [59] and endorsed in later decisions of this court including Tan Cheng Bock v Attorney-General [2017] 2 SLR 850 (“Tan Cheng Bock”) at [37] (in the context of interpreting the Constitution) and Public Prosecutor v Lam Leng Hung and others [2018] 1 SLR 659 (“Lam Leng Hung”) at [67] (in the context of interpreting the Penal Code). This framework requires the court, first, to ascertain the possible interpretations of the text in question; second, to ascertain the legislative purpose of the statute; and third, to compare the possible interpretations of the text against the ascertained legislative purpose.
76 The respondent submitted, at the first step, that there were two possible interpretations of the expression “a child above 7 years of age and under 12” in s 83: first, that it referred to persons whose chronological age was between seven and 12, as well as persons whose chronological age was above 12 but whose mental age was less than 12; and second, that it referred only to persons whose chronological age was between seven and 12. The respondent contended that this ambiguity existed mainly because the Penal Code did not define the words “age” and “child”, and because adopting the first interpretation of the expression was not inconsistent with other provisions in the Penal Code. In our judgment, however, on a proper understanding of the ordinary meaning of the word “age” in s 83, there was no ambiguity of the kind suggested by the respondent. Accordingly, there was no need to undertake the second and third steps of the three-step analysis set out in Ting Choon Meng in order to determine which of the possible interpretations of the expression ought to be preferred in the light of the purpose of s 83 and that of the Penal Code as a whole.
77 As we observed in Tan Cheng Bock, the possible interpretations of a statutory provision are ascertained by “determining the ordinary meaning of [its] words” (at [38]). One aspect of the principle that words in a legislative enactment are to be given their ordinary meaning is that words mean what they were understood to mean at the time they were adopted by the Legislature. There is, after all, no other objective basis upon which to construe the meaning which the framers of the legislation intended the legislative text to have.
78 Perhaps the clearest recent example of this approach is our decision in Lam Leng Hung. The question in that case was whether a director of a company was an “agent” within the meaning of s 409 of the Penal Code, which created the offence of criminal breach of trust by an agent. We answered the question in the negative, and one of our reasons for this was that at the time the Penal Code was adopted, the operative expression in s 409, namely, “in the way of his business as … an agent”, referred to a person who was engaged in the trade of being a professional agent, such as a property or insurance agent, as opposed to simply any person who was regarded by the law as being in a relationship of agency to another person who would be regarded as his principal. This was borne out, we found, partly by the kinds of agent which s 409 specifically enumerated, such as bankers, attorneys and factors, and partly by applying the interpretive maxim that where a general term follows a number of specific terms of an identifiable class, the general term is to be construed as if it were qualified by the features of that class. Commenting on the meaning of the word “factor” in s 409, what was significant to the court was the ordinary meaning which that word bore at the time the Penal Code was drafted. We said at [148]:
The crucial point for present purposes is that while the factor originally occupied “an important but subordinate position as a species of servant”, by the 18th century the factor “came to assume an independent status, often of great financial strength, buying and selling on commission on behalf of his various principals” … ([see Roderick Munday, “A Legal History of the Factor” (1977) 6 Anglo-Am LR 221] at 259). In the circumstances, we find that there is little historical basis to support the Prosecution’s submission that a factor is merely a species of an employee or clerk. On the contrary, the historical material demonstrates that by the time the Penal Code was drafted, a factor was regarded as a commercial tradesman engaged in the trade or profession of dealing with foreign merchants on behalf of his principal, and who simultaneously also carried out a number of other important financial functions such as giving advances to his principal and granting credit to purchasers. [emphasis in original omitted; emphasis added in italics]
79 In the present case, therefore, to determine the ordinary meaning of the word “age” in s 83, the court must begin by considering what that word was understood to mean at the time the Penal Code was adopted, that is, in 1872. In this regard, it is significant that at that time, the concept of mental age had not yet been developed or articulated. As we have seen at [55] above, Alfred Binet came up with this concept only in 1905. Therefore, when the Penal Code (including s 83) was first enacted, nobody in the Straits Settlements, including the Legislative Council, would have thought that “age” in s 83 referred to or included mental age. The only other possible meaning of “age” is “a period of history”, such as in the expressions “the Stone age” or “the Elizabethan age”. But the context of s 83, which refers to offenders “above 7 years of age and under 12”, shows that this is not a reasonable alternative. There is therefore no ambiguity in the ordinary meaning of the word “age” in s 83, which must mean chronological age.
80 Be that as it may, since the concept of mental age did not exist at the time the Penal Code was adopted, it is arguably also not possible to say that s 83 was not intended to refer to that concept, given that such an intention could not have been formed at that time in the first place. By way of analogy, if the question in Lam Leng Hung had been whether a Bitcoin broker was an “agent” within the meaning of s 409, it would not have sufficed to say that because Bitcoin brokers did not exist at the time the Penal Code was adopted, the Legislature could not have intended s 409 to apply to them, and therefore, they are not caught by s 409. Instead, the court would have had to consider whether the ordinary meaning and concept of the word “agent” in s 409 at the time the Penal Code was enacted could logically extend to Bitcoin brokers. And the court might have concluded, for example, that because a Bitcoin broker is fundamentally a kind of professional agent, he would be caught by s 409.
81 In short, where the court is considering new phenomena which did not exist at the time the provision in question was adopted, it must consider whether the ordinary meaning of the provision at that time can logically extend to the new phenomena. Here, we considered that the ordinary meaning of the word “age” in s 83 did not logically extend to the subsequently developed concept of mental age. The former is fundamentally a measure of time, specifically, the length of time that a person has lived since birth, whereas the latter is not. It is instead a measure of a person’s cognitive ability with reference to a specific mathematical formula: see [50] above. The two concepts are therefore qualitatively different, and cannot be reduced to a common denominator in the way, for example, that a Bitcoin broker and a factor might both be regarded as professional agents for the purposes of s 409. In our judgment, therefore, the word “age” in s 83 does not include mental age.
82 Other textual elements of s 83 and the Penal Code support this view. First, s 83 applies to a “child” who is between seven and 12 years of age. It would be odd if s 83 could be relied upon by a 40-year-old with a mental age of between seven and 12 because such a person would be an “adult”, and not a “child”. Second, as the Prosecution observed, the word “age” is used in other provisions of the Penal Code to refer to chronological age, and adopting that interpretation of the same word in s 83 would be consistent with the general presumption that where an expression is used in a statute in different instances, it should bear the same meaning: see Tan Cheng Bock at [58(c)(i)]. Thus, for example, s 310 states that it is an offence for a woman to cause the death of her child, “being a child under the age of 12 months”, and s 377D(2) defines a “minor” as a person “under 21 years of age”. The word “age” in these provisions clearly refers to chronological age, and barring any contraindication, “age” in s 83 ought to be interpreted in the same way.
83 The respondent was able to cite one authority, State v Mazid (25 April 2015) (Delhi District Court) (“Mazid”), to support his interpretation of s 83. In that case, the accused, who had an IQ of 70 and a mental age of ten and a half years, faced a charge of kidnapping. It is not clear from the judgment what his chronological age was. The court acquitted him of the charge on the basis of s 83 of the Indian Penal Code (Act 45 of 1860) (“the IPC”), which is in pari materia with s 83 of the Penal Code. In particular, the court appeared to consider that the word “age” in s 83 of the IPC included mental age because that concept was established by scientific opinion. The court explained its reasoning in these terms (at [27]–[30] per Dr PS Malik):
27. There is one more factor in this case and that is the mental age of the accused was about 10 ½ years. It shows that he possessed only that much of mental calibre as a child of 10 ½ years usually appears to possess at this age.
28. This Court as [sic] already perused, as aforesaid, the language of Seciton [sic] 83 IPC which specifically lays down that nothing is an offence which is done by a child above seven years of age and under twelve, who has not attained sufficient maturity of understanding to judge of the nature and consequences of his conduct on that occasion.
29. This factum of the mental age of [the] accused as to have been 10 ½ years has been completely approved by medical practitioners of a specialist body i.e. IHBAS and it cannot be doubted in its intention or competence. On this account also the accused appears to be having no bad intention i.e. mens rea required to commit the offence of murder of the kidnapped child or kidnapping of the child as aforesaid.
30. With these observations, this Court is of the view that there is nothing on record to show the culpability of the commission of [the] offence as is charged against the accused. He is accordingly acquitted of the charge framed against him.
84 We were unpersuaded by this reasoning. No attempt was made in Mazid to discern the ordinary meaning of the word “age” in s 83 of the IPC. It appears that the court thought that importing the concept of mental age into the provision was a salutary idea given the acceptance of the concept by medical practitioners, and that this was sufficient basis for reading the word “age” in the provision as including that concept. In our judgment, this did not reflect a proper approach to the interpretation of legislation, at least not in this jurisdiction. Therefore, even though Mazid reached a conclusion which supports the respondent’s view that s 83 includes the concept of mental age, we gained no assistance from its reasoning as to why s 83 ought to be so interpreted. In the circumstances, we held that the word “age” in s 83 refers to chronological age and not mental age.
The constitutionality of s 83
85 On the hypothesis that s 83 did not apply to him, the respondent submitted that the provision was contrary to Arts 9(1) and 12(1) of the Constitution. We did not agree.
86 We begin with Art 12(1), which provides: “All persons are equal before the law and entitled to the equal protection of the law.” The test that is applied to analyse whether a legislative provision is consistent with an individual’s right to equal protection under Art 12(1) is known as the “reasonable classification” test. This test requires that the classification prescribed by the provision be founded on an intelligible differentia, and that the differentia bear a rational relation to the object sought to be achieved by the statute: see Lim Meng Suang and another v Attorney-General and another appeal and another matter [2015] 1 SLR 26 (“Lim Meng Suang”) at [60]. The reasonable classification test, generally speaking, is not difficult to satisfy. As we noted in Lim Meng Suang, “very seldom” [emphasis in original omitted] would a legislative provision contain differentia that is not intelligible (at [65]), and there is also no need for a “perfect relation” or “complete coincidence” [emphasis in original omitted] between the differentia and the statute’s object in order for there to be a rational relation (at [68]).
87 In this case, the impugned differentia in s 83 is chronological age. Specifically, it is the characteristic of being “above 7 years of age and under 12”, in the words of the provision. The respondent submitted that this differentia bore no rational relation to the object sought to be achieved by s 83, which, in the respondent’s view, was “to excuse apparent culpability by taking into account the maturity of the offender”. According to the respondent, persons of the same chronological age might have different levels of maturity of understanding, and the differentia of chronological age failed to take into account those who were mentally challenged, resulting in their being treated worse than persons who were of similar maturity but who had the requisite chronological age. Accordingly, s 83 was said to be inconsistent with Art 12(1).
88 In our judgment, this argument was misconceived. In the first place, we considered, accepting the Prosecution’s submission, that the purpose of s 83 is broader than that contended by the respondent. It is partly to excuse young children from criminal liability because they are likely to have very low culpability for their offending acts in view of their incomplete intellectual development, and also partly to protect them from the harshness of the criminal justice system: see Singapore Parliamentary Debates, Official Report (21 February 2000) vol 71 at cols 937–939 (Abdullah Tarmugi, Minister for Community Development); Stanley Yeo, Criminal Law for the 21st Century: A Model Code for Singapore (Academy Publishing, 2013) at paras 7.1.2 and 7.1.10. Bearing this purpose in mind, we considered that a chronological age range of seven to 12 years is plainly a reasonable criterion for delineating the class of persons to whom the protection in s 83 should apply. It is a meaningful proxy for identifying persons who are likely to require that protection because the younger a person, the less mature he is likely to be. The criterion is neither perfect nor accurate, but the reasonable classification test does not require that. Indeed, similar reasoning was employed by this court in Yong Vui Kong v Public Prosecutor [2015] 2 SLR 1129 (“Yong Vui Kong (caning)”) to explain why it was not irrational for legislation to preclude men above fifty from being caned (at [116]):
In our judgment, the use of age as a convenient proxy to screen out those who are likely to be unfit for caning is plainly reasonable and passes muster under the second limb of the reasonable classification test. This is because there is an inverse relationship between one’s age and one’s physical condition. That this differentia might be over-inclusive (in that some males over the age of 50 might still be fit for caning) is not fatal, since – as we have stated earlier – there is no need for a perfect coincidence between the differentia used and the object sought to be achieved. …
89 Next is Art 9(1), which provides: “No person shall be deprived of his life or personal liberty save in accordance with law.” The question raised is whether s 83 counts as “law” within the meaning of Art 9(1). It is well established that there are two ways of showing that a particular provision is not law in this sense. The first is to show that it is inconsistent with a higher law in Singapore and is therefore not law. For example, it has been argued, unsuccessfully, that legislative provisions establishing caning as a form of punishment are not “law” under Art 9(1) because they are inconsistent with a higher law prohibiting torture: see Yong Vui Kong (caning). No argument of this nature was pursued by the respondent. The second way is to show that the alleged law is so arbitrary and absurd that it does not constitute “law” under Art 9(1): see Prabagaran a/l Srivijayan v Public Prosecutor [2017] 1 SLR 173 (“Prabagaran”) at [90]; Yong Vui Kong v Public Prosecutor and another matter [2010] 3 SLR 489 at [16]. In Prabagaran, we observed at [93] that this involves an inquiry that is “in substance, no different from that under Art 12(1) of the Constitution”. Given our conclusion that s 83 is consistent with Art 12(1), there was no merit to the respondent’s Art 9(1) challenge against s 83.
90 We were satisfied, therefore, that s 83 is consistent with Arts 9(1) and 12(1).
The relevance to sentencing of s 83 and its neighbouring provisions
91 Although we considered that s 83 did not apply and that, as drafted, it is consistent with the Constitution, we thought that s 83 and its neighbouring provisions shed some light, at a general level, on the Legislature’s attitude towards how the culpability of young offenders should be analysed. The relevant provisions read as follows:
Act of a child under 7 years of age
82. Nothing is an offence which is done by a child under 7 years of age.
Act of a child above 7 and under 12 years of age, who has not sufficient maturity of understanding
83. Nothing is an offence which is done by a child above 7 years of age and under 12, who has not attained sufficient maturity of understanding to judge of the nature and consequence of his conduct on that occasion.
Act of a person of unsound mind
84. Nothing is an offence which is done by a person who, at the time of doing it, by reason of unsoundness of mind, is incapable of knowing the nature of the act, or that he is doing what is either wrong or contrary to law.
92 It seemed to us that these provisions contain a general philosophy that the younger an offender, the less serious the diminishment of his mental capacity needs to be in order to provide a defence to his crime and, logically flowing from this, also to reduce his culpability for sentencing purposes even where the defence may not be available. Thus, while nothing is an offence which is done by a child under seven, nothing is an offence which is done by a child between seven and 12 only if he has not attained sufficient maturity of understanding to judge the nature and consequence of his offending conduct; and for everyone else who is above the age of 12, nothing is an offence only if the offender, at the time of the relevant act, by reason of unsoundness of mind, is incapable of knowing the nature of the act, or that he is doing what is either wrong or contrary to law, this being a stricter standard compared to the previous, and even more so compared to the first. The corollary is that as a general proposition, the older an offender, the more serious the diminishment of mental capacity that must be proved in order for him to be exculpated and, following from this as a matter of logical deduction, also for his culpability to be reduced for sentencing purposes where the defence is not available. We illustrate the point with an example: suppose that an offender is 12 years and a day old. Clearly, the defence in s 83 would not apply. But it seems to us that it would take quite little, and probably much less than would be the case if he were 25 years old, to say that his culpability is significantly diminished if it can be shown that, by reason of his age, he lacks maturity of understanding.
93 Of course, close attention to the facts of the case remains necessary to ascertain the degree of a young offender’s culpability. But the point is that it can be seen even in the substantive criminal law that the Legislature has contemplated a general correlation between youth and folly; and in our view, this is further support not only for the appropriateness of using mental age as a heuristic tool for analysing culpability, but also for both the imperative reflected in our sentencing jurisprudence that the young should be given a second chance as well as the sentencing principles which have been built upon that conviction (which are discussed at [122] below).
Issue 3: The appropriate sentencing framework
94 We turn next to the appropriate framework for sentencing intellectually disabled young offenders who have been convicted of serious offences. In brief, we considered that the two-step framework articulated in Al-Ansari ([39] above) for sentencing young offenders for serious offences was suited to the task, albeit with a minor qualification to a gloss placed on that framework by the High Court in Boaz Koh ([39] above). To illustrate the need for this qualification, we consider it useful to begin with some general observations on the nature of the exercise of sentencing young offenders.
95 That exercise, when serious offences are concerned, may be said to possess two uncommon features. First, in sentencing a young offender for a serious offence, the court often has a relatively wide range of sentencing options at its disposal, and must choose between them. These include probation and reformative training; any punishment which the offence in question provides for, whether it be imprisonment, caning, fine or a combination of them; and also community sentences where appropriate. There is therefore the need in every such case to reason out which of these qualitatively different sentencing options is most appropriate. This is unlike the usual case of sentencing an adult offender, where the task of the sentencing court typically is to impose an appropriate sentence within the statutorily prescribed range of punishments. The second uncommon feature is that rehabilitation is presumed to be the dominant sentencing objective for young offenders unless otherwise shown: see Public Prosecutor v Mok Ping Wuen Maurice [1998] 3 SLR(R) 439 at [21]. This is a reflection of: (a) young offenders’ generally lower culpability due to their immaturity; (b) their enhanced prospects of rehabilitation; (c) society’s interest in rehabilitating them; and (d) the recognition that the prison environment may have a corrupting influence on young offenders, who are more impressionable and susceptible to bad influence than older offenders: see Sundaresh Menon CJ, “Keynote Address at the Sentencing Conference 2017” (26 October 2017) at paras 19–21; see also A Karthik v Public Prosecutor [2018] 5 SLR 1289 (“Karthik”) at [37]–[42].
96 Naturally, the second feature has an effect on the first feature, in the sense that if rehabilitation is established as the dominant sentencing objective, then the choice of sentencing option has to be guided by that objective. The Al-Ansari framework is fundamentally built on a recognition of this logical relationship. That is why it articulates a two-step framework under which the court, at the first step, considers whether rehabilitation ought to be the dominant sentencing objective, and, at the second step, chooses the appropriate sentencing option in the light of the answer at the first step. V K Rajah JA put it in this way in Al-Ansari at [77]–[78]:
77 Accordingly, in dealing with sentencing young offenders involved in serious offences, I propose the following analytical framework. First, the court must ask itself whether rehabilitation can remain a predominant consideration. If the offence was particularly heinous or the offender has a long history of offending, then reform and rehabilitation may not even be possible or relevant, notwithstanding the youth of the offender. In this case, the statutorily prescribed punishment (in most cases a term of imprisonment) will be appropriate.
78 However, if the principle of rehabilitation is considered to be relevant as a dominant sentencing consideration, the next question is how to give effect to this. In this respect, with young offenders, the courts may generally choose between probation and reformative training. The courts have to realise that each represents a different fulcrum in the balance between rehabilitation and deterrence. In seeking to achieve the proper balance, the courts could consider the factors I enumerated above [(at [67])], but must, above all, pay heed to the conceptual basis for rehabilitation and deterrence.
97 In Boaz Koh, the High Court endorsed this two-stage analysis and, with reference to the first step, discussed the circumstances in which rehabilitation might be displaced as the dominant sentencing consideration. The court observed as follows at [30]:
… [R]ehabilitation is neither singular nor unyielding. The focus on rehabilitation can be diminished or even eclipsed by such considerations as deterrence or retribution where the circumstances warrant. Broadly speaking, this happens in cases where (a) the offence is serious, (b) the harm caused is severe, (c) the offender is hardened and recalcitrant, or (d) the conditions do not exist to make rehabilitative sentencing options such as probation or reformative training viable.
98 In the present case, the Prosecution relied on all four factors, placing particular emphasis on factor (d), to persuade us that rehabilitation had been displaced as the dominant sentencing consideration. Specifically, the thrust of its factual case on appeal, as was its case in the court below, was that the respondent was too intellectually disabled to undergo reformative training, which required a certain minimum level of cognitive ability. Given that the conditions which made reformative training viable did not exist, rehabilitation, in the Prosecution’s view, could not be the dominant sentencing consideration.
99 In our judgment, while this argument appeared to be consistent with Boaz Koh, it was inconsistent with the logic of the two-step framework in Al-Ansari which we have described above. The question whether it is desirable that an offender be rehabilitated must be conceptually distinguished from the question whether he is suitable for reformative training. It is the former, and not the latter, which determines whether rehabilitation should be the dominant sentencing consideration in his case. The reason for this is that an offender’s suitability for reformative training indicates only whether he is suitable to undergo a specific form of rehabilitation. It does not indicate whether normatively, he should be rehabilitated, in the sense that it would be in society’s best interests that rehabilitation be the controlling sentencing objective. That is the issue at the first step of the Al-Ansari framework. At that stage, the court is not yet concerned with the operational question of how rehabilitation ought to be achieved. The court will certainly have to grapple with that question eventually, but to leap to it directly is to place the cart before the horse. In short, the existence of practical constraints on achieving rehabilitation which are external to the offender does not entail that he should not be rehabilitated. The existence of such constraints properly influences the process of deciding the appropriate sentencing option, and not the process of deciding whether rehabilitation should be the dominant sentencing objective.
100 The consequences of ignoring the distinction mentioned above are significant. If, for example, rehabilitation is jettisoned as a relevant sentencing consideration at the first step of the analysis on the basis that it would be difficult to implement, then when the court considers the sentencing options at the second stage, some other sentencing consideration, such as deterrence or incapacitation, would assume dominance, and the court would resolve to choose a sentencing option which gives effect to that. If, however, rehabilitation is normatively established as the dominant sentencing consideration at the first step regardless of the challenges in its implementation, then the court would be driven to choose a sentencing option that gives effect to it notwithstanding those challenges. In our judgment, this is the right approach where it is desirable that the offender be rehabilitated notwithstanding practical constraints external to him which present difficulties for the rehabilitative process.
101 On this view, the seeming placement in Boaz Koh of factor (d) at the first step of the Al-Ansari framework must be clarified. In that case, the High Court placed factor (d) at the first step to address a particular kind of case, namely, where the young offender was a foreign national who was not locally resident. The court gave the example of Long Yan v Public Prosecutor Magistrate’s Appeal No 9015 of 2015 (16 July 2015) (“Long Yan”), noting that the young offender in that case “was a foreign national and … had no family in Singapore”: see Boaz Koh at [33]. Such circumstances give rise to at least two special considerations. First, the offender may have to be repatriated after serving his sentence, and if so, there would be no instrumental reason for rehabilitating him since his leaving the country would already fulfil the ultimate aim at which his rehabilitation would have been directed, namely, to protect society from his possible future crimes. Second, existing rehabilitative options would generally be unworkable. Probation is unlikely to be feasible because the offender has no kin to supervise his conduct: see Public Prosecutor v Fernando Payagala Waduge Malitha Kumar [2007] 2 SLR(R) 334 (“Fernando”) at [17]; see also Tan Choon Huat v Public Prosecutor [1991] 1 SLR(R) 863 at [24]. For the same reason, reformative training is unlikely to be suitable because it likewise involves supervision after incarceration: see Public Prosecutor v Lai Leng Hwa and another [1991] 2 SLR(R) 214 at [25]. Therefore, where the offender’s circumstances are similar to those of the offender in Long Yan, it is sensible that rehabilitation will generally not be a relevant sentencing consideration from the outset. That said, in other circumstances, rehabilitation may well apply to an offender who is a foreign national, as the case law has recognised: see Fernando at [17]. All will depend on the facts of the particular case at hand.
102 In cases which do not involve foreign offenders who are not locally resident, however, factor (d), in our judgment, properly falls under the second step of the Al-Ansari framework. That is because the question whether the conditions exist to make a certain rehabilitative option viable is fundamentally an operational question of how rehabilitation might be achieved, which is the question addressed at the second step. In the present case, this means that even if the Prosecution were right to say that reformative training as it is currently designed is not suitable for the respondent by reason of his intellectual disability, this does not mean that rehabilitation has been displaced as the normative sentencing consideration at the first step of the Al-Ansari framework. To persuade us of that displacement, the Prosecution must instead provide positive reasons as to why sentencing considerations other than rehabilitation are dominant.
Issue 4: Applying the first step of the sentencing framework
103 This leads us to the first step of the Al-Ansari framework. Whether rehabilitation was displaced as the dominant sentencing consideration in this case turned principally on the respondent’s state of mind at the time of his offences. His state of mind at that time made it clear to us that deterrence was of reduced significance. Although it also indicated that he posed a high risk of reoffending, his youth and his mental impairment pointed to rehabilitation, and not incapacitation, as the preferred crime prevention objective. Retribution was relevant as well, but in a different way, in that it is an abiding consideration in sentencing that the sentence imposed must be proportionate to the gravity of the offence. In this case, the gravity of the respondent’s offences was significantly attenuated by his reduced culpability, and this was an important consideration at the second step of the Al-Ansari framework, as will be seen.
The respondent’s state of mind at the material time
104 The respondent’s state of mind at the time of his offences shed light on his culpability, the kind of offender he is and his risk of reoffending. To develop a proper view on this issue, we found it useful to distinguish five concepts that were somewhat loosely employed by both sides, namely:
(a) the knowledge which the respondent had by virtue of the nature of the offences of aggravated rape and sexual assault by penetration, these being the offences which were the subject of the charges proceeded with;
(b) the respondent’s intellectual disability;
(c) the causal link between the respondent’s intellectual disability and his offending acts;
(d) the respondent’s knowledge of the legal rightness or wrongfulness of his offending acts; and
(e) the respondent’s knowledge of the moral rightness or wrongfulness of his offending acts.
105 First, the knowledge which the respondent had by virtue of the nature of the offences of aggravated rape and sexual assault by penetration refers to knowledge of the kind that an offender must have had when committing these offences. In this regard, the Prosecution submitted that “[t]he nature of aggravated rape requires the [r]espondent to know that he penetrated the [v]ictim’s vagina with his penis without her consent and put her in fear of hurt while doing so” [emphasis in original], and that “[t]he nature of sexual assault by penetration requires the [r]espondent to know that he penetrated the [v]ictim’s vagina with his finger, and a comb, without her consent” [emphasis in original]. This argument did not seem entirely accurate to us because these two offences do not “require” the offender to “know” that the victim did not consent. They require only that he did not believe in good faith that the victim consented: see Pram Nair v Public Prosecutor [2017] 2 SLR 1015 (“Pram Nair”) at [110]–[111] and s 79 of the Penal Code. But even if the Prosecution were right in submitting that the knowledge which it referred to inhered in the nature of these two offences, this too was not helpful because it did not assist us in differentiating between the different degrees of culpability of different offenders who commit these offences.
106 Second, the respondent’s intellectual disability concerned the broader question of his overall intellectual functioning at the time of his offences. Plainly, this was distinct from any knowledge which he might have had by virtue of the nature of his offences. As we have seen, the respondent was assessed to have an IQ of 61, which meant that his cognitive ability was equal to or better than just 0.5% of his same-aged peers, and he had a mental age of between eight and ten at the time of his offences: see [49] and [67] above. Unlike a psychiatric disorder, which may manifest episodically and give the afflicted person moments of lucidity in between, an intellectual disability is by nature a permanent condition which affects the person at every moment. On that basis, the respondent’s intellectual disability was a prima facie reason to regard his culpability as reduced because, as a general matter, he does not understand the world around him as well as the average person of his chronological age. For that reason, it was sensible to consider the respondent as prima facie belonging to a class of offenders who should be treated less severely. This also meant that evidence could be produced to show that he ought not to be regarded as such.
107 Third, the existence of a causal link between the respondent’s intellectual disability and his offending acts represented a specific means by which his intellectual disability reduced his culpability, namely, by affecting his control over his offending impulses. In that sense, the concept of causal link related to the process by which the respondent’s mens rea was formed and his willingness to act on his offending impulses. It was for that reason also distinct from his knowledge of the legal and moral rightness or wrongfulness of his conduct. This distinction was reflected in Dr Winslow’s and Dr Sarkar’s opinion that while the respondent knew that what he had done to the victim was wrong, his low IQ and poor adaptive behaviour likely contributed to his inability properly to control his criminal impulses: see [69] above. As we have observed, Dr Cai agreed with this opinion, and the Prosecution accepted it: see [70]–[72] above.
108 Fourth, the respondent’s knowledge of the legal rightness or wrongfulness of his actions concerned the specific issue of whether the respondent knew that the law prohibited his offending acts. This was relevant to assessing his culpability because while ignorance of the law is not a defence to criminal liability, it may in certain circumstances suggest that an offender did not have the intention of breaking the law, and is therefore less culpable. This is likely to be relevant where the offender’s failure to apprehend the illegality of the acts in question stems from his lack of mental maturity and understanding. For this reason, we rejected the Prosecution’s submission that the respondent’s understanding of the legal rightness or wrongfulness of his actions was irrelevant for the purposes of assessing his culpability.
109 In this case, there was no evidence that the respondent understood at the time of his offences that what he had done to the victim was legally wrong. The Prosecution highlighted the fact that a month before the respondent committed the aggravated rape and the sexual assault by penetration offences, he had been arrested for outraging the modesty of a 21-year-old girl by grabbing her buttocks, and submitted that since he had given a cautioned statement in respect of that charge, it would have been brought home to him that touching a girl in a sexual manner without her consent was wrong. We found it difficult to give weight to this fact. First, at the time of the aggravated rape and the sexual assault by penetration offences, the respondent had not been convicted of the outrage of modesty charge. He had simply been accused of having committed that offence, with no judgment made as to his guilt. Therefore, it would be unfair to draw the conclusion that he should have known from his arrest for that offence that he had broken the law on that occasion, and was again breaking the law when he raped and sexually assaulted the victim. Second, there is no evidence as to what the respondent registered from this experience of being arrested. His low IQ might have impeded his appreciating and remembering that he had been accused of breaking the law. Accordingly, we found nothing on record to show that the respondent deliberately broke the law when he raped and sexually assaulted the victim.
110 Fifth, the respondent’s knowledge of the moral rightness or wrongfulness of his actions concerned the question of whether he knew that his offending acts were wrong for any reason. Here, the evidence was clear that the respondent knew that what he had done to the victim was wrong. The question was the extent of this awareness. Based on the responses which the respondent gave Dr Lee when the latter asked him about the incident in September 2015, almost a year after he committed the aggravated rape and the sexual assault by penetration offences, we did not think he fully appreciated the implications of his actions on the victim. Dr Lee’s report dated 20 October 2015 stated:
55. On direct questioning on why he said that he will take out knife, he replied:
- “Because I just say any how. I don’t know. For fun”.
56. On direct questioning if he had previously threatened people before, he said:
- “No, I didn’t know about anything. So I think of something and do to her”.
57. On direct questioning on why he said “I love you baby”, he replied:
- “I watched video movie”
58. On direct questioning on the possible result of having sex with the alleged victim, he said:
- “Make baby. Become family”.
59. When asked if he wanted to become a family with the alleged victim, he replied:
- “I not sure. If her parents agree, I accept”.
- “If her parents didn’t want, can’t do anything”.
60. When asked why he had inserted a comb into the vagina of the alleged victim, he said:
- “Comb? I put inside vagina. I was searching from something. I forget what thing. I suddenly think of something to put inside vagina. I feel very angry, must poke vagina”.
61. When questioned further on his act of poking the comb into the vagina, he said that the girl could feel pain, but he said that he did not want the girl to feel pain. When asked to provide reason for him inserting the comb, he said:
- “I never think, I nothing to do, I don’t know. I find comb, feel angry, want to poke vagina. Now I regret. I feel sad, I do this thing to girl, now the girl will never want to marry me or love me”.
- “I feel sad. I poke girl, she still virgin. I feel sad, I think I this [sic] to my younger sister, elder sister, then no good”.
- “That day I feel very bad, put comb in vagina, when I kanna caught I think back”.
[underlining and emphasis in italics in original]
111 The Prosecution did not challenge these statements. The impression that we had from reading them was of a person with a distorted, confused but ultimately simplistic view of sexuality, and of the significance and heinousness of the sexual abuse that he committed. The respondent was recorded as having experienced a sense of remorse which arose out of a regret that the victim would not reciprocate his feelings. Critically, he appeared not to have even begun to understand the depravity of his conduct, the degradation and trauma suffered by the victim, and the consequences for the both of them. Unsurprisingly, Dr Lee concluded in his report that the respondent knew that his actions were wrong, but “did not appreciate the legal wrongfulness of his act[s] because of his mental retardation (defect in the mind) and lack of appropriate prior instructions concerning consensual sex”.
112 Against this lay the Prosecution’s suggestion that the respondent appeared to know enough about the wrongfulness of his actions to the extent that he “elected to stalk [the victim] for 15 whole minutes until they were alone in a secluded corner”, accosted her in the absence of witnesses, and fled the scene after committing the offences. While this might be true, it did not tell us anything about what the respondent thought was wrong about his actions. By contrast, Dr Lee testified that the respondent’s sense of right and wrong came from a fear of his mother, and the Prosecution did not challenge this testimony. In our judgment, the fact that the respondent regarded the wrongfulness of his actions as arising from the fact that he had disobeyed his mother, and not from the nature of his conduct, was yet another sign of how inadequately he appreciated the true meaning and implications of what he had done, and this in turn diminished his culpability.
113 To conclude, we thought that it was evident that the respondent’s cognitive ability was extremely low, and that this significantly reduced his culpability. His intellectual disability compromised his ability to control his impulses. He also manifested a limited understanding of the nature and consequence of his actions. A reasonable person would be able to appreciate that aggravated rape and sexual assault by penetration are serious offences involving a profound violation of the victim’s dignity, privacy and bodily integrity, and result in both immediate distress and long-term trauma. But, as we have seen, the respondent did not understand that what he had done was of this nature and consequence: see [110]–[112] above. Likely contributing to his inability to judge the nature and consequence of his actions was his ignorance of their illegality and his simplistic view of their moral wrongfulness. In the circumstances, we considered that his culpability for his offences was substantially reduced. Coupled with his youth, this placed the focus on rehabilitation as the dominant sentencing objective.
Deterrence
114 In that light, we turn to explain why we did not think that rehabilitation was displaced by any other sentencing objective as the dominant sentencing objective, beginning with deterrence.
115 In our judgment, the extent of the respondent’s intellectual disability significantly reduced the importance of both general and specific deterrence in this case. As we observed in Soh Meiyun v Public Prosecutor [2014] 3 SLR 299 at [43], general deterrence is premised on the cognitive normalcy of both the offender in question and the potential offenders sought to be deterred: see also Public Prosecutor v Kong Peng Yee [2018] 2 SLR 295 (“Kong Peng Yee”) at [69]. Thus, the precise weight to be accorded to general deterrence would depend on, among other things, the causal link between the offender’s intellectual disability and the offence: see Kong Peng Yee at [70]. Specific deterrence assumes that the offender can weigh the consequences before committing an offence. It is therefore unlikely to be effective when the offender’s ability fully to appreciate the nature and quality of his actions is reduced: see Kong Peng Yee at [72]. As we have seen, the respondent is not cognitively normal, and did not fully understand the gravity of his offending conduct. Deterrence in both forms must therefore carry minimal weight here.
116 This view is supported by the decision of the High Court of Australia in Muldrock v The Queen (2011) 244 CLR 120 (“Muldrock”). The appellant in that case, a 30-year-old man with a mild intellectual disability, pleaded guilty to an offence of sexual intercourse with a boy aged under ten. He had befriended the boy, and had taken advantage of an opportunity when the two were alone swimming to fellate the boy. He was sentenced by the New South Wales District Court, which found him to be “significantly intellectually disabled” (at [10]), to nine years’ imprisonment for his offence, with a non-parole period of 96 days. The New South Wales Court of Criminal Appeal allowed the Prosecution’s appeal against the sentence by increasing the non-parole period to six years and eight months. The appellant appealed to the High Court of Australia, arguing that the New South Wales Court of Criminal Appeal had wrongly rejected the New South Wales District Court’s finding that he was intellectually disabled, and as a result, had failed to give that proper weight.
117 The evidence showed that the appellant had been sexually abused as a child, and that his intellectual disability was the cause of his difficulty in managing his impulses and controlling his actions (at [39]–[40]). He had an IQ of 62, and functioned at a level lower than 99% of the population. His receptive and expressive language was equivalent to that of a child aged five and a half years, and he functioned in the lowest 0.1% of the population in terms of his adaptive behaviour (at [42]). A psychologist who assessed him considered that he would benefit from a programme designed for a sex offender with an intellectual disability (at [44]). She suggested that he needed to learn practical skills for dealing with situations in which he was near children, and that he needed a comprehensive programme to address the areas of deficit in his adaptive behaviour (at [44]).
118 The High Court of Australia allowed the appellant’s appeal, holding that the original sentence of nine years’ imprisonment was manifestly excessive. It restored the New South Wales District Court’s finding that the appellant had a significant intellectual disability, and remitted the case back to the New South Wales Court of Criminal Appeal for the appellant to be sentenced with proper weight given to his intellectual disability. In this regard, the court opined that the assessment that the appellant suffered from a “mild intellectual disability” should not obscure the fact that he was mentally retarded because, in the scientific literature, “significantly subaverage intellectual function” was defined as an IQ of about 70 or below (at [50]). The court then set out the principles for sentencing intellectually disabled offenders, and applied them as follows (at [53]–[55]):
53 … One purpose of sentencing is to deter others who might be minded to offend as the offender has done. Young CJ, in a passage that has been frequently cited, said this [(R v Mooney (unreported, Court of Criminal Appeal (Vic), 21 June 1978) at 5)]:
“General deterrence should often be given very little weight in the case of an offender suffering from a mental disorder or abnormality because such an offender is not an appropriate medium for making an example to others.”
In the same case, Lush J explained the reason for the principle in this way [(at 8)]:
“[The] significance [of general deterrence] in a particular case will, however, at least usually be related to the kindred concept of retribution or punishment in which is involved an element of instinctive appreciation of the appropriateness of the sentence to the case. A sentence imposed with deterrence in view will not be acceptable if its retributive effect on the offender is felt to be inappropriate to his situation and to the needs of the community.”
54 The principle is well recognised. It applies in sentencing offenders suffering from mental illness, and those with an intellectual handicap. A question will often arise as to the causal relation, if any, between an offender’s mental illness and the commission of the offence. Such a question is less likely to arise in sentencing a mentally retarded offender because the lack of capacity to reason, as an ordinary person might, as to the wrongfulness of the conduct will, in most cases, substantially lessen the offender’s moral culpability for the offence. The retributive effect and denunciatory aspect of a sentence that is appropriate to a person of ordinary capacity will often be inappropriate to the situation of a mentally retarded offender and to the needs of the community.
55 In this case, there was unchallenged evidence of the causal relation between the appellant’s retardation and his offending in the reports of Dr Muir and Ms Daniels. The fact that the appellant possessed the superficial understanding of a mentally retarded adult that it was wrong to engage in sexual contact with a child and that he told childish lies in the hope of shifting the blame from himself were not reasons to assess his criminality as significant, much less to use him as a medium by which to deter others from offending.
[emphasis added]
119 Similarly, in this case, the respondent’s state of mind made him quite unsuitable, in our judgment, to be used as a medium by which to deter others from offending. His inability to appreciate the full significance of what he had done also made it difficult for us to give any weight to specific deterrence.
Incapacitation
120 We turn next to the sentencing objective of incapacitation. The Prosecution’s case on appeal was based largely on persuading us that prevention, in the form of incapacitation, was more important than rehabilitation in this case. The Prosecution emphasised that we had to be mindful of the need to protect the public. It pointed to the respondent’s inability to control his offending impulses and his slowness to reform due to his intellectual disability, arguing that these factors indicated that he presented a high risk of recidivism. During oral argument, the Prosecution furthered this submission by repeatedly emphasising that the court was fundamentally engaged in a “risk assessment” exercise. In this regard, the Prosecution posed the question: “What is the risk appetite that the court has?”, and implored us to be guided by the need to choose the “less risky option”.
121 In our judgment, this was the wrong approach. By framing the issue as one which involved fundamentally a risk assessment exercise, the Prosecution made two errors of principle. First, it failed to have regard to the fact that the respondent is a young offender for whom rehabilitation is, as a matter of principle, the dominant sentencing consideration by default. In this regard, the Prosecution failed to grapple with the issue of principle as to how, if at all, the existence of the “risk” which it was fixated upon justified incapacitation over rehabilitation as the appropriate crime prevention objective. As we shall explain, incapacitation generally does not trump rehabilitation where young offenders are concerned. Second, the Prosecution’s approach failed to have regard to the need for proportionate sentencing. We address this in the next section of these grounds.
122 In Karthik ([95] above), the High Court explained that there were two primary reasons why young offenders should be sentenced on the basis of rehabilitation being the dominant sentencing consideration (at [37]). The first is that a young offender should be given a second chance because of his youthful folly and inexperience. This reason is retrospective in nature, being based on the offender’s age at the time of the offence (at [37(a)]). The second is that rehabilitation is the preferred tool for discouraging young offenders from future offending because such offenders are, by reason of their youth, more amenable to reform, and society would benefit considerably from their rehabilitation. Further, young offenders are unduly affected compared to adult offenders when they are exposed to typical punitive sentencing options such as imprisonment. This second reason is prospective in nature, being based on the offender’s age at the time of sentencing (at [37(b)]).
123 In this case, the Prosecution’s emphasis on the need to manage the respondent’s risk of reoffending did little to engage the above concerns. In particular, it did not explain why the respondent ought not to be given a second chance in view of his youthful folly. We reiterate that the respondent was only 14 at the time of his offences, and had a mental age of between eight and ten then. The retrospective reason for rehabilitation therefore applied in this case, indeed, with particular force. The Prosecution also failed to address the possibility that incapacitating the respondent as a method of managing his risk of reoffending might turn out to be counterproductive, both for the respondent and for society. After all, the respondent was not even 18 at the time of sentencing, and thus, there was no reason why the prospective rationale for rehabilitation should not apply as well. Yet, before us, the Prosecution failed, just as it did before the Judge, to address that concern, which was eloquently expressed by the Judge as follows (see the GD at [101]):
… [T]he pivotal question that remained unanswered was what conclusion should be drawn from the premise that the Accused’s risk of recidivism was significant? In this regard, I did not agree with the Prosecution that a long period of incarceration would necessarily be the panacea to the risk of recidivism. Indeed, with respect, the submission that incarceration would protect the public appeared to me to be short-sighted on the facts of this case. At the time of sentencing, the Accused was 17 years of age. Even discounting remission and backdating, and even if the Accused was sentenced to 18 years’ imprisonment which was the highest end of the Prosecution’s sentencing position, what would become of the Accused and of those around him when he is subsequently released in his early thirties? Would society be better protected when the Accused is released from incarceration, stronger and bigger, but lacking insight into the consequences that his choices and conduct carry?
124 We agreed with these observations, and were therefore not persuaded that incapacitation had displaced rehabilitation as the appropriate crime prevention sentencing objective in this case.
125 Potentially inconsistent with the approach set out above is the High Court’s decision in Iskandar bin Muhamad Nordin v Public Prosecutor [2006] 1 SLR(R) 265 (“Iskandar”). The accused in that case was an 18-year-old male with an IQ of 58. He pleaded guilty to a charge of outrage of modesty under s 354 of the Penal Code for grabbing a stranger’s breast. The district judge sentenced him to nine months’ imprisonment and three strokes of the cane, and he appealed. The High Court held that the weight to be attached to the accused’s intellectual disability depended on the particular circumstances of the case (at [9] and [14]). On the facts, Yong Pung How CJ considered that it did not attract mitigating weight because, in his view, it “did not impair [the accused’s] ability to gain insight into his actions as well as the consequences of those actions” (at [18]). Yong CJ dismissed the appeal and increased the accused’s sentence to 24 months’ imprisonment and nine strokes of the cane on the basis that “[his] unrepentant and aberrant behaviour showed that he posed a real danger to the community”, and that “[a] longer term of imprisonment would represent a longer period of protection for society from his depredations” (at [22]).
126 While we agreed that the significance of an offender’s intellectual disability to the sentencing exercise would depend on the facts of the case, we were troubled that even though the accused in Iskandar was only 18, and therefore, a young offender, the court considered incapacitation as being on an equal, if not greater, footing than rehabilitation. We did not think that this was right, even in the light of the accused’s antecedents, which comprised two prior convictions for theft, resulting in two weeks’ and six weeks’ imprisonment respectively. With a young offender, rehabilitation should be preferred to incapacitation as the applicable crime prevention sentencing objective. If the intention in Iskandar had been to incarcerate the accused for a period of two years, it might have been better to sentence him to reformative training.
127 In any case, as we indicated to the Prosecution during oral argument, the prevention of crime could not be pursued at any cost. Important to the calibration of the proper sentence for this purpose is the sentencing objective of retribution and its quantitative proxy, proportionality, to which we shall now turn. We shall consider this in some detail because, in our judgment, it was crucial to articulate an applicable method of analysing proportionality in order properly to assess the suitability of the vastly disparate proposed sentences in this case.
Retribution
128 The principle of retribution holds that the punishment imposed should reflect the degree of harm that has been occasioned by the offence and the offender’s culpability in committing it: see Public Prosecutor v Loqmanul Hakim bin Buang [2007] 4 SLR(R) 753 at [46]. Harm is generally measured by looking at the consequences of the offence, for example, the degree of trauma and degradation suffered by the victim in this case. Culpability, on the other hand, is generally assessed by looking at the offender’s state of mind when he committed the offence, for example, whether he committed the offence with premeditation, out of rashness or under an afflicted mind.
129 The essence of the concept of retribution is the relationship between criminal punishment and the past crime for which it is imposed, rather than any contingent future benefits or objectives which such punishment might bring about or help to secure: see R A Duff, “Retrieving Retributivism” in Mark D White, Retributivism: Essays on Theory and Policy (Oxford University Press 2011) ch 1 at p 3. Retribution justifies punishment by reference to the rightness alone of punishment as an institution, and not by the beneficial consequences which that institution is thought to generate, such as the prevention of harm to society through the rehabilitation or incapacitation of the offender, or the deterring of future offending. It is not the task of the sentencing court to reconcile at a philosophical level divergent theories of punishment, and in any event, it has been recognised that “any morally tolerable account of [the institution of criminal punishment] must exhibit it as a compromise between distinct and partly conflicting principles”: see H L A Hart, “Prolegomenon to the Principles of Punishment” in Punishment and Responsibility ([1] above) ch 1 at p 1. It is, however, the court’s task to reason within that compromise by giving effect to each of the classical sentencing objectives under a coherent and rational framework, and in a manner sensitive to the facts of each case.
130 Critical to that task, in our judgment, is a proper appreciation of how each sentencing objective features in the criminal process and, consequently, in the reasoning process of the sentencing court. Sentencing objectives that focus on the outcomes or consequences for the offender, namely, deterrence, incapacitation and rehabilitation, are associated with guiding the court towards a sentence which achieves the outcome that is regarded as the most desirable on the facts of the case at hand. When one outcome is, for some reason, preferred to the others, that outcome is commonly called the dominant sentencing consideration, and often, there is a dispute over which outcome befits that label. By contrast, retribution is an objective that exerts a much broader influence on the criminal process: the general requirement that both mens rea and actus reus must be present, the phenomenon that only the offender is punished for his crime, and the principle that his punishment should be proportionate to his crime are all retributive ideas. Together, they ensure that the criminal sanction which is imposed is an appropriate response in terms of expressing society’s denunciation of the offence.
131 Retribution does not easily lend itself to being treated as a dominant sentencing objective. It requires that the sentence imposed be commensurate with the offender’s culpability and the harm that he has caused. Certainly, on this metric, the greater the degree of either, the more severe the punishment should be. Where less severe punishment is imposed because the degree of culpability and harm is small, retribution does not lose its significance. It simply operates to justify the less severe sentence. In this sense, retribution is generally an abiding objective during sentencing. An example of this phenomenon is the fact that judge-made sentencing frameworks for a variety of statutory offences take the common form of a harm-culpability matrix which embeds in a fundamental way retributive thinking in the sentencing exercise: see, for example, Poh Boon Kiat v Public Prosecutor [2014] 4 SLR 892 at [77]–[78] (vice offences under the Women’s Charter (Cap 353, 2009 Rev Ed)); Public Prosecutor v Tan Thian Earn [2016] 3 SLR 269 at [28] (possessing a controlled substance useful for the manufacture of methamphetamine under s 10A(1)(c) of the Misuse of Drugs Act (Cap 185, 2008 Rev Ed)); Public Prosecutor v Lim Yee Hua and another appeal [2018] 3 SLR 1106 at [27] (voluntarily causing hurt under s 323 of the Penal Code in the context of “road rage” cases); Public Prosecutor v Yeo Ek Boon Jeffrey and another matter [2018] 3 SLR 1080 at [59] (voluntarily causing hurt to a public servant under s 332 of the Penal Code).
132 That said, it is possible for retribution to be “displaced” as a sentencing consideration to some degree where the court is satisfied that the punishment should not correspond completely to the offender’s culpability and the harm that he has caused. For example, a sentence potentially less severe than what the offender deserves may be imposed because of his strong rehabilitative prospects. That is why probation is a sentencing option for young offenders and offenders who have committed offences whose relative lack of seriousness is reflected through the absence of a sentence fixed by law: see s 5(1) of the Probation of Offenders Act (Cap 252, 1985 Rev Ed) (“the POA”). On the other hand, a sentence potentially more severe than what the offender deserves may be imposed to protect the public. An example of this is Kong Peng Yee ([115] above), where the offender was convicted of culpable homicide not amounting to murder for killing his wife during a psychiatric episode (which the evidence showed was unlikely to recur). We allowed the Prosecution’s appeal against the two-year imprisonment sentence imposed by the High Court, and imposed a longer sentence of six years’ imprisonment instead, “not to punish him but to try to achieve the twin objectives of rehabilitation and prevention (resulting in the protection of others) in the best way possible” [emphasis added] (at [96]).
133 Part of the reason why retribution may be “displaced” in this way is because it is hard to say with absolute quantitative precision what an offender deserves for his crime. That is why, where a sentencing framework is in play, the general practice is that the court first identifies the appropriate sentencing range on the basis of the harm caused and the offender’s culpability, and then, having regard to other sentencing factors, including outcome-focused sentencing objectives, chooses the appropriate sentence within that range. The range operates as a margin of reasonableness which ensures that the eventual sentence imposed remains broadly proportionate to the crime. However, in a case such as Kong Peng Yee where outcome-focused sentencing objectives are particularly compelling, we think it is critical also to apply with special rigour the totality principle enunciated in Mohamed Shouffee bin Adam v Public Prosecutor [2014] 2 SLR 998 (“Shouffee”). As was said in Shouffee at [47], this principle is “a manifestation of the requirement of proportionality that runs through the gamut of sentencing decisions”. Two of its requirements are that: (a) the aggregate sentence generally should not be substantially above the normal level of sentences for the most serious of the individual offences committed; and (b) the aggregate sentence must not be crushing or out of step with the offender’s past record and future prospects: see Shouffee at [54] and [57]. In our judgment, it is these principles, coupled with a careful application of the relevant sentencing framework, that enables the court in each case to strike the proper balance between retributive and outcome-focused sentencing considerations, and to say that all things considered, “the punishment fits the crime”: see Public Prosecutor v Kwong Kok Hing [2008] 2 SLR(R) 684 at [42]. And it was in this light that we assessed the aggregate sentence sought in this case, as we shall elaborate below. With this in mind, we turn now to the second step of the Al-Ansari analysis, namely, the determination of the appropriate sentencing option in view of our finding, at the first step of the analysis, that rehabilitation was the dominant sentencing consideration.
Issue 5: Applying the second step of the sentencing framework
134 There were three sentencing options available in law for the respondent: first, a probation order; second, reformative training; and third, imprisonment with caning. Under the third option, the court would have to sentence the respondent to more than eight years’ imprisonment and at least 12 strokes of the cane. This is because aggravated rape carries a mandatory minimum sentence of eight years’ imprisonment and 12 strokes of the cane under s 375(3)(a) of the Penal Code; and pursuant to s 307(1) of the CPC, as the respondent was convicted at one trial of three distinct offences, the court was obliged to order the sentences for at least two of the offences to run consecutively. Below, we explain, firstly, why we considered that rehabilitation, being the dominant sentencing objective, entailed that reformative training was the only justifiable sentence in principle, and, secondly, why the Prosecution’s proposed sentence of between 15 and 18 years’ imprisonment and at least 15 strokes of the cane was disproportionate and thus unjustified.
The sentencing options available
Probation
135 Not even the respondent asked the court to make a probation order, and for good reason. Probation would not sufficiently recognise the seriousness of the respondent’s offences: see Public Prosecutor v Muhammad Nuzaihan bin Kamal Luddin [1999] 3 SLR(R) 653 at [16]. Nor would it sufficiently recognise that the respondent continues to pose a threat to society because of his high risk of recidivism. Probation would do little to provide the structured environment which is necessary for that risk to be managed and treated with the assurance that it will not manifest itself in antisocial or otherwise harmful behaviour towards others during the period of rehabilitation. To the extent that a probation order may contain a residence requirement (see s 5(3)(a) of the POA), s 5(3A) of the POA states that the period of residence in an approved institution (as defined in s 2 of the POA) cannot extend beyond 12 months from the date of the order, which, on any estimation, is clearly too short a period in this case. Therefore, the Judge correctly considered probation to be inappropriate: see the GD at [34].
Imprisonment with caning
136 Next, we considered that imprisonment with caning was also precluded as a matter of principle because it was not an option which gave primary effect to rehabilitation as a sentencing consideration. On this point, we agreed with the Judge: see the GD at [136]. As to imprisonment, it has been “readily acknowledge[d] that a term of standard imprisonment cannot be said to place the principle of rehabilitation as a dominant sentencing consideration”: see Al-Ansari at [65]. And as to caning, it is well established that it is imposed as a deterrent punishment: see Amin bin Abdullah v Public Prosecutor [2017] 5 SLR 904 at [62]. Deterrence had a minimal role to play in this case, and retribution required the sentence imposed to be constrained within proportionate bounds.
137 The Prosecution submitted that a lengthy imprisonment sentence with caning was the “best option” for “both the [r]espondent and society at large”. In addition to its incapacitating effect, such a sentence, the Prosecution argued, would give the authorities more time to rehabilitate the respondent. In this regard, the Prosecution relied on the testimony of Mr Ong Pee Eng (see [23] above) that the SPS was able to provide adequate care for intellectually disabled offenders by adjusting their method of instruction and arranging appropriate cell mates for them. In addition, Mr Ong testified that the SPS was committed to rehabilitating the respondent, and would be prepared to arrange for him to be taught on a one-to-one basis and given individualised general psychological intervention. The Prosecution added that the respondent would be able to continue his rehabilitation under the Conditional Remission System upon his release, or could be placed on the Mandatory Aftercare Scheme to facilitate his reintegration into society.
138 We were not persuaded by these arguments. If it was accepted that in sentencing the respondent, deterrence was not applicable because of his circumstances, and incapacitation was outweighed by rehabilitation for the same reason, then it would be wrong in principle to require the respondent to suffer a sentence which furthered objectives that had no rational application to him in this case.
139 The Prosecution also did not address the Judge’s observation that although the SPS is able to provide customised rehabilitative treatment for the respondent, no arguments were advanced to explain why he could not receive the same in the context of reformative training. Both regimes are run by the SPS. The Prosecution came close to addressing this point when it submitted that even if customised rehabilitative treatment were provided, reformative training would still be carried out using the same content as that designed for people of normal intelligence; and because group-learning would be difficult for the respondent, reformative training “can only use an incomplete syllabus to educate him” [emphasis in original omitted]. The Prosecution distinguished the case of Mohammad Fadlee, an example which Mr Ng himself cited and which the Judge regarded as showing that “suitable modifications” could be made to adjust reformative training for intellectually disabled offenders (see [23(d)] above), on the basis that the offender in that case had a “substantially higher” [emphasis in original omitted] level of intellect (an IQ of 70–79), and was assessed to be mentally fit for reformative training.
140 We found these submissions difficult to accept. The court’s task in this case was not to attempt to identify the best form of rehabilitation for the respondent from a therapeutic perspective. It was clear from the evidence of both Dr Cai and Dr Lee that the best form of rehabilitative treatment which the respondent could receive was specialised treatment catering to his intellectual disability in a mental facility. The court could only take as its starting point the fact that such specialised treatment unfortunately cannot be provided in existing state-run rehabilitation facilities, and had to decide how, within those parameters, the respondent might best be rehabilitated, given that rehabilitation was the dominant sentencing consideration in this case. The fact that customised measures would be imperfect was only to be expected, and did not lead to the conclusion that no such measures were worth pursuing if they were the only option which the court had that was supported by principle.
141 In that respect, the Prosecution, in our view, missed the point when it submitted that it was not proper for the SPS to be “compelled” to customise programmes for offenders like the respondent because this would be a policy decision for the Executive, and not the Judiciary. All that the Judge was doing was to discern the limits of what can be provided during reformative training in order to decide whether the applicable sentencing considerations in this case could be adequately given effect through such a sentence. And that, in our judgment, was an entirely appropriate exercise, not only for the Judge, but also for this court.
142 Finally, the Prosecution submitted that the Judge erred in being concerned not to expose the respondent to the corrupting influence of the prison environment because the respondent had already been keeping bad company and abusive habits, and therefore, “[w]hatever the perils of prison, it is clear that the [r]espondent already has no shortage of such ‘corrupting influence’ in his environment out of prison”. We did not hesitate to reject this argument. If one were genuinely concerned to reduce the respondent’s risk of recidivism, then surely, the conscionable approach would be to protect him from further corrupting influence, and not to tolerate his further exposure to it – indeed, for 15 to 18 years – on the supposition that he has already been somewhat exposed to it.
Reformative training
143 The Prosecution submitted that reformative training was unsuitable because it was designed “solely” to rehabilitate offenders of normal intelligence as its curriculum was based on cognitive behavioural theory. The Prosecution relied on Dr Rajesh’s opinion that the respondent’s low verbal comprehension ability would prevent him from participating in group learning exercises, and that his weak working memory suggested that he would find it difficult to use the information from the reformative training programme to avoid further criminal behaviour. The Prosecution also relied on the concurring views of Mr Soh Tee Peng William (see [23] above) and Dr Cai, both of whom were pessimistic about the respondent’s prospects of rehabilitation through reformative training.
144 We did not agree. We did not accept that rehabilitation in a reformative training centre would not work for the respondent. There is no justification in the CPC as to why a suitable programme should not exist for an offender like him. If there were a need for any law reform in this regard, then that would be a matter for the Legislature. Moreover, although the respondent has an intellectual disability, the evidence suggested that he is not wholly devoid of cognition, and is capable of reasoning at a very basic level. That, combined with the progress he is recorded to have made at the Singapore Boys’ Home, showed that he has some capacity for rehabilitation which could not be ignored in determining whether he should be sent for reformative training: see the GD at [172]–[173].
145 We also made two observations on the Prosecution’s submission that reformative training as it is currently designed would not be effective, given that the respondent lacked the ability to understand what had caused him to commit his offences or to understand what he needed to do in order to help prevent a recurrence of his offending conduct. First, this submission emphasised precisely why the respondent’s culpability should be viewed as significantly diminished in the circumstances. Second, it highlighted the potential injustice and inequality that would arise if the benefits of reformative training were to be denied to intellectually disabled offenders when it seems that they have particular need for such training. There was force in the respondent’s argument that if a decision not to impose reformative training were made on the basis that his intellectual disability rendered him unsuitable for it, this might be seen as excluding an entire class of persons – the intellectually disabled – from benefiting from reformative training. This much was recognised by the Judge too: see the GD at [154]. And, in our judgment, it was no answer to that objection for the Prosecution to say, as it did, that even if the respondent did not have an intellectual disability, it would have sought the same sentence. If the respondent did not have that disability, it seems inevitable that the analysis of his culpability would have been entirely different, as would likely have been the relevance of other sentencing objectives such as deterrence.
Proportionality
146 Having explained why we concluded that reformative training was the only sentencing option that could be justified as a matter of principle, we turn now to complete the analysis by explaining why we considered that the quantum of the sentence proposed by the Prosecution would, in any event, have been disproportionate.
147 The Prosecution’s position was that a sentence of between 15 and 18 years’ imprisonment and at least 15 strokes of the cane should be imposed on the respondent. The Prosecution submitted that his offence of aggravated rape fell under Band 2 of the sentencing framework set out in Ng Kean Meng Terence v Public Prosecutor [2017] 2 SLR 449 (“Terence Ng”) for rape offences. Band 2 of that framework prescribes an indicative sentencing range of 13 to 17 years’ imprisonment and 12 strokes of the cane: see Terence Ng at [47(b)]. The Prosecution also submitted that the respondent’s two offences of sexual assault by penetration fell under Band 2 of the sentencing framework set out in Pram Nair ([105] above) for that offence. Band 2 of that framework prescribes an indicative sentencing range of ten to 15 years’ imprisonment and eight strokes of the cane: see Pram Nair at [159(b)]. In view of the totality principle, the Prosecution proposed the global sentence mentioned above.
148 In our judgment, the Prosecution’s position failed to account for the fact that the respondent’s culpability was substantially reduced because of the extent of his intellectual disability. The respondent has an intellectual disability which diminished his ability to control his impulses; there is no evidence that he knew he was breaking the law at the time of his offences; and his understanding of the wrongfulness of his conduct was distorted and simplistic: see [106]–[113] above.
149 The Terence Ng framework, in our judgment, was not crafted with this sort of offender in mind; nor was the Pram Nair framework. This is evident from the reasoning of the court in both cases. In Terence Ng, the offender-specific mitigating factors were contemplated to justify a reduction only within an indicative sentencing band, and they were said to include remorse, youth and advanced age. The court did not consider how an offender with exceptionally reduced culpability because of cognitive impairment would fit within that framework. Nor did we do so in Pram Nair, where we adapted the Terence Ng framework to the offence of sexual assault by penetration. It seems to us that even if Terence Ng or Pram Nair could be extrapolated to an offender with the respondent’s mental characteristics, the starting point would be a sentence below Band 1 of either framework.
150 For this reason, we also considered that the sentence sought by the Prosecution would have violated both limbs of the totality principle. As to the first limb of this principle, the most serious offence in this case is the offence of aggravated rape. A sentence below Band 1 of the Terence Ng framework for this offence would have been a sentence of between eight and ten years’ imprisonment, eight years being the mandatory minimum imprisonment term, and at least 12 strokes of the cane (see s 375(3) of the Penal Code), and this would have been the normal range of sentences for the offence: see Haliffie bin Mamat v Public Prosecutor and other appeals [2016] 5 SLR 636 at [79]. This range is considerably lower than the aggregate sentence of between 15 and 18 years’ imprisonment and at least 15 strokes of the cane suggested by the Prosecution, indicating that the latter was disproportionate. In our judgment, the Prosecution’s proposed sentence would also have been crushing on the respondent and not in keeping with his past record and future prospects under the second limb of the totality principle. Specifically, such a sentence would have been a dramatic leap from any punishment which he had ever received for his early juvenile offending, and we have no reason to believe that it would result in his rehabilitation, as we have noted at [123]–[124] above.
151 But we did not only find the Prosecution’s proposed sentence disproportionate. We considered also, and perhaps more importantly, that reformative training was a proportionate sentence in this case, having regard to the sentences imposed for offences committed in similar circumstances. For the purposes of this comparison, there did not appear to be any local case involving facts similar to those in this case, so we found guidance from a number of foreign decisions.
152 In R v Taylor (Ross Gordon) (1983) 5 Cr App R (S) 241 (“Taylor”), a decision of the English Court of Appeal which was cited by the respondent, the accused, a 21-year-old intellectually disabled man, pleaded guilty to the rape of a 19-year-old girl with Down’s Syndrome. They attended the same school, and had met on the way home on the day of the offence. As they approached an alley, the accused forced himself on the girl, but desisted after realising that she was having her period. He was sentenced to three years’ imprisonment, and he appealed. The English Court of Appeal allowed his appeal and substituted a three-year probation order. It noted that the psychiatric evidence showed that the accused’s sex education and understanding of his own sexuality was limited, and that he had a limited capacity to understand the full implications of what had occurred (at 242). We accept that there are aspects of the court’s reasoning which seem problematic, including an observation that the act which the accused committed was not so much a rape but more in the nature of an “indecent assault” (at 243). But apart from that, the court did incline to the view that the accused had to be regarded as less culpable because he did not fully appreciate his wrongdoing. Peter Pain J said at 243:
One then has difficulties about the mental equipment of this appellant. It is quite apparent from the papers and from what he said to the police and so on that he knew that what he was doing was wrong. But it is not sufficient in this sort of case to say that he knew the difference between right and wrong, and that that is the end of the matter. Where a person has an almost childlike understanding, as it is plain this appellant has – and indeed it has to be borne in mind that he was still attending school – one has to appreciate that such a person may understand that something is wrong without appreciating just how very wrong it is, just as a child may understand that something is wrong without understanding that it is very wicked. We are inclined to think that that was the position here. He did not appreciate at the time – and the various reports seem to make this clear – that what he was doing was not just something that he ought not to be doing, but that it was something that was very wicked indeed. [emphasis added]
153 Like the accused in Taylor, the respondent in this case had but a childlike understanding of the wholly inexcusable, depraved and destructive conduct that he had engaged in. And just as the court in Taylor did not think that imprisonment was the appropriate sentence, so too we considered that the respondent did not deserve a long prison term and caning. Similar reasoning was adopted in MBQ and Muldrock, discussed above at [61] and [116]–[118] respectively. In both cases, the offenders committed serious sexual offences: MBQ involved the rape of a three-year-old toddler, and Muldrock involved the fellatio of a young boy. But due to the offenders’ substantially reduced culpability, lengthy incarceration was considered inappropriate in both cases. Taylor, MBQ and Muldrock, in our judgment, indicate that a sentence of reformative training in this case could not be criticised as being manifestly inadequate or otherwise out of line and therefore inconsistent with the sort of sentences that might be imposed for broadly similar cases.
154 Against this conclusion lay the Prosecution’s submission during oral argument that sentencing the respondent to reformative training would create a “schism” between how the law treats intellectually disabled offenders who have committed serious offences and how the law treats mentally disordered offenders who have committed serious offences. The Prosecution contended that in the latter type of case, the approach was always to regard the “protection of the public” as the dominant sentencing consideration. For this proposition, the Prosecution cited Lim Ghim Peow v Public Prosecutor [2014] 4 SLR 1287 (“Lim Ghim Peow”) and Public Prosecutor v Goh Lee Yin and another appeal [2008] 1 SLR(R) 824 (“Goh Lee Yin”). As there was no meaningful distinction between intellectual disability and mental disorder for the purposes of reducing culpability, the approach in relation to mentally disordered offenders who committed serious offences, the Prosecution contended, would be thrown in doubt if the court in this case did not similarly impose a lengthy incapacitating sentence on the respondent.
155 In our judgment, this submission was misconceived on at least three levels. First, this case concerned a young offender, whereas Lim Ghim Peow and Goh Lee Yin did not. The special priority attaching to rehabilitation as a sentencing objective therefore did not apply in those two cases. Second, it did not assist the Prosecution to emphasise the importance of protecting the public because rehabilitation, like incapacitation, is also a method of achieving that goal: see Kong Peng Yee at [78]. And as we have observed at [123] above, the Prosecution could not adequately explain why incapacitation in this case was the superior method. Third, the Prosecution’s submission contained no attempt to explain on a principled basis how the imposition of an effective incapacitating sentence ought to be squared with exacting only punishment that is deserved and proportionate to the offender’s crime. As we suggested to the Prosecution during oral argument, if the protection of the public were so important, why not impose the maximum term of imprisonment?
156 The authorities relied on by the Prosecution certainly did not support the imposition of disproportionate sentences on offenders in order to protect the public. For example, in Lim Ghim Peow, the accused, a 46-year-old man, pleaded guilty to a charge under s 304(a) of the Penal Code of culpable homicide not amounting to murder for killing his ex-lover over their broken relationship. He had doused her with petrol which he had prepared the night before and set her ablaze. At that time, he was suffering from a major depressive disorder. The High Court imposed a sentence of 20 years’ imprisonment, and the Court of Appeal affirmed this. Although the accused’s depression was found to have limited his perception of the choices available to him, it was held that “there was nothing to indicate that [he] lacked the capacity to comprehend his actions or appreciate the wrongfulness of his conduct” (at [52]). Lim Ghim Peow was therefore not a case in which the offender’s state of mind was regarded by the court as having substantially reduced his culpability. In other words, it was not a case where the offender received an incapacitating sentence that he did not deserve. It was consistent with the proposition that the punishment imposed must, save possibly in exceptional circumstances, remain proportionate to the crime.
157 Lim Ghim Peow may be contrasted with Public Prosecutor v Aniza bte Essa [2009] 3 SLR(R) 327. In that case, the accused, a 24-year-old woman, pleaded guilty to a charge under s 304(a) read with s 109 of the Penal Code of abetting culpable homicide not amounting to murder for obsessively urging her lover to kill her husband, which he eventually did. At the time of her offence, the accused was suffering from moderate depression. The High Court imposed a sentence of nine years’ imprisonment, and the Court of Appeal affirmed this, rejecting the Prosecution’s appeal to impose a sentence of life imprisonment instead. The psychiatric evidence showed that the accused’s depression was of moderate severity at the time of the offence, concurrent with her having received increasingly frequent beatings from her husband, and that she also showed some features of post-traumatic stress disorder (at [8]). There was “a complex interplay between her chronic depression, and her constantly being abused”, which “resulted in her failure to act positively to prevent [her lover] from following through with the plan to kill the deceased” (at [8]). On these facts, this court approved the High Court’s view that the accused’s culpability “was not … so high that she deserved the sentence of life imprisonment” (at [64]). Nor did this court think that the accused’s risk of reoffending ought to be suppressed with a sentence disproportionate to her culpability. Chan Sek Keong CJ put it in this way (at [40]):
… [I]n our view, to sentence a mentally unstable offender (whose condition is treatable) to life imprisonment, because at that point of time we do not know with certainty when it is safe to release him or her back to society, seems to be unjust to such an offender. It would mean punishing such an offender out of proportion to his or her culpability. The burden is on the Prosecution to satisfy the court that such a treatable offender is likely to remain a danger to the community if he or she is released back to society (see [Public Prosecutor v Chee Cheong Hin Constance [2006] 2 SLR(R) 707] at [16]). In our view, the Judge’s approach could not be faulted. The fault, if any, lay in the then existing limitations of a penal regime that required the Judge to impose either a sentence which might be (and which he considered) excessive or a sentence which might be seen to be inadequate (eg, by the [Public Prosecutor]). [emphasis in original]
158 We agreed with these observations, and, in particular, with the view expressed in the last sentence of the passage above that limitations in the sentencing regime are no justification for disproportionate sentencing. In this case, the Prosecution’s position appeared to be that if the court did not have the tools to rehabilitate the respondent, then it should punish him. We could not accept this. It is for either the Executive or Legislature to create the appropriate tools, and to the extent that they have not, that omission is not a principled basis for exacting disproportionate punishment on an offender.
Conclusion
159 In the end, we were faced with a choice between imposing a term of between 15 and 18 years’ imprisonment and at least 15 strokes of the cane, and imposing a term of incarceration at a reformative training centre of up to three years. In our judgment, if those were both sub-optimal options, then in all the circumstances, the latter was the less imperfect and only principled option, especially considering that the respondent had already been incarcerated for almost four years. The sentence of reformative training could not and would not be backdated, and in all the circumstances, we were satisfied that this was the appropriate course.
160 For all these reasons, we dismissed the appeal.
Sundaresh Menon Andrew Phang Boon Leong Judith Prakash
Chief Justice Judge of Appeal Judge of Appeal
Chief Justice
Tay Yong Kwang Chao Hick Tin
Judge of Appeal Senior Judge
Judge of Appeal
Kow Keng Siong, Sarah Shi, Carene Poh and Jamie Pang (Attorney-General’s Chambers) for the appellant;
N Sreenivasan SC, Jerrie Tan and Jason Lim (Straits Law Practice LLC), Amarick Gill (Amarick Gill LLC), Muntaz Zainuddin (IRB Law LLP) and Cheryl Ng (Intelleigen Legal LLC) for
the respondent.
SUPREME COURT OF SINGAPORE
11 March 2019
Case summary
Public Prosecutor v ASR [2019] SGCA 16
Criminal Appeal No 10 of 2018
----------------------------------------------------------------------------------------------------------------
Decision of the Court of Appeal (delivered by Chief Justice Sundaresh Menon):
CoA upholds sentence of reformative training for intellectually disabled teenager convicted of serious sexual offences and clarifies applicable sentencing principles.
Background
-
In 2014, the respondent raped the victim after threatening to pull a knife on her; he also inserted his finger and a blunt object into her vagina. He was 14 at that time and had an intellectual disability, with an IQ of 61 and a mental age of between eight and ten. For his actions, he was charged under the Penal Code (Cap 224, 2008 Rev Ed) with one count of aggravated rape and two counts of sexual assault by penetration. In 2017, after he had turned 16, he pleaded guilty to the three charges and consented to having six other charges taken into consideration for sentencing purposes. The High Court sentenced him to reformative training: at [9], [10], [12]–[14] and [28].
-
The Prosecution appealed, maintaining its position below that the respondent should be sentenced to between 15 and 18 years’ imprisonment and at least 15 strokes of the cane. It submitted that the High Court had wrongly held that rehabilitation was the dominant sentencing objective, mainly because the respondent’s offences were serious, the harm caused was severe, the respondent was recalcitrant and his intellectual disability made him unsuitable for reformative training. The Prosecution contended that a long imprisonment term was the only suitable sentencing option, principally because the respondent had a high risk of recidivism, and the public needed to be protected from him: at [32]–[33].
-
The respondent defended the High Court’s decision, arguing that the benefit of reformative training should not be denied to intellectually disabled persons such as him. Reformative training was suitable because it would be in both his and society’s long-term interests that he be rehabilitated, and because his culpability, and thus the seriousness of his offences, had been reduced by his intellectual disability: at [35].
-
The parties also addressed the effect of s 83 of the Penal Code, which provides an absolute defence to criminal liability for any child “above 7 years of age and under 12” who was not of sufficient maturity of understanding to judge the nature and consequences of his offending conduct at the time of such conduct. The Prosecution submitted that s 83 did not apply because the word “age” in the provision referred to an offender’s chronological age rather than his mental age, and the respondent was above 12 at the time of his offences. Furthermore, the differentia of chronological age did not render the provision unconstitutional. The respondent, in contrast, contended that s 83 ought to be read purposively to apply to offenders whose mental age was between seven and 12 regardless of whether their chronological age was above or below 12; and if that were not the case, then the provision would be unconstitutional for unfairly excluding the benefit of the defence from those of insufficient maturity by reason only of their mental age: at [29]–[31] and [34].
The Court of Appeal’s decision
-
The Court of Appeal (“the Court”) dismissed the appeal after hearing the parties. In its detailed grounds of decision, the Court set out its reasoning on five broad issues which had arisen: (a) the extent of the respondent’s intellectual disability; (b) the applicability and constitutionality of s 83 of the Penal Code; (c) the appropriate sentencing framework for intellectually disabled young offenders convicted of serious offences; (d) whether rehabilitation was the dominant sentencing objective in this case; and (e) in the light of the applicable sentencing objectives in this case, whether reformative training was the appropriate sentencing option: at [36]–[41].
Issue 1: The extent of the respondent’s intellectual disability
-
The Court considered that the extent of the respondent’s intellectual disability centred around the significance of his IQ, the degree to which his intellectual disability affected his ability to control his impulses, and the significance of his mental age. As to his IQ, the evidence supported the view that at the time of his offences, he functioned in the extremely low range of intelligence. As to his impulse control, the evidence was clear, and the Prosecution accepted, that his intellectual disability impaired his impulse control ability: at [43], [49] and [68]–[72].
-
The concept of mental age, in the Court’s view, was not heterodox, outdated or unreliable, contrary to the Prosecution’s submission. Instead, it was useful as a heuristic tool for approximating the intellectual capacity of an offender and, in turn, assessing his culpability for sentencing purposes. It simply needed to be understood in the context of the offender’s life experiences and circumstances. This view was supported by the cases. Here, it was meaningful to consider that the respondent had a mental age of between eight and ten at the time of his offences. There was no evidence that he had by then accumulated life experiences and participated in activities which would suggest that he was more mature than his mental age indicated: at [50], [51], [57], [58], [63], [66] and [67].
Issue 2: Section 83 of the Penal Code
-
The Court held that there was no ambiguity in the ordinary meaning of the word “age” in s 83 of the Penal Code that needed to be resolved by purposive interpretation. One aspect of giving words in a legislative provision their ordinary meaning was that the words in question meant what they were understood to mean at the time they were adopted by the Legislature. When the Penal Code was adopted, the concept of mental age had not been developed, so “age” in s 83 could not mean mental age. Nor did the concept of chronological age logically extend to mental age. Moreover, other textual elements of s 83 and the Penal Code supported this interpretation. The respondent, being above 12 at the time of his offences, therefore could not rely on s 83: at [76]–[84].
-
This reading of s 83 did not make the provision unconstitutional. It did not violate the respondent’s right to equal protection under Art 12(1) of the Constitution because the differentia of chronological age bore a rational relation to the provision’s objective of excusing young children from criminal liability and protecting them from the harshness of the criminal justice system. Nor did s 83 violate Art 9(1) by being inconsistent with any higher law or being so arbitrary and absurd that it did not constitute law: at [85]–[90].
Issue 3: The appropriate sentencing framework
-
The Court held that the appropriate sentencing framework for intellectually disabled young offenders convicted of serious offences was the two-step approach laid down by the High Court in Public Prosecutor v Mohammad Al-Ansari bin Basri [2008] 1 SLR(R) 449. This consisted of, first, considering whether rehabilitation remained the dominant sentencing consideration (it being presumptively so in the case of young offenders), and, second, choosing the appropriate sentencing option in the light of the answer at the first step. As a general rule, whether conditions existed to make rehabilitative sentencing options viable was a factor properly considered at the second step because it concerned the operational question of how rehabilitation might be achieved, and not the normative question of whether it would be in society’s best interests that the offender be rehabilitated. In this case, therefore, the respondent’s alleged unsuitability for reformative training by reason of his intellectual disability was not a reason for displacing rehabilitation as the presumptive dominant sentencing objective, contrary to the Prosecution’s submission: at [94], [96], [98], [99] and [102].
Issue 4: Whether rehabilitation was displaced
-
Whether rehabilitation was displaced as the dominant sentencing consideration turned largely on the respondent’s state of mind at the time of his offences. In the Court’s view, the evidence showed that the respondent was of extremely low cognitive ability, and this significantly reduced his culpability. His intellectual disability impaired his impulse control ability, and he manifested a limited understanding of the nature and consequences of his offences as well as their legal and moral wrongfulness. Coupled with his youth, this made rehabilitation the dominant sentencing objective: at [103] and [106]–[113].
-
Rehabilitation was not displaced from this position by other sentencing objectives. Deterrence carried minimal weight because it was predicated in part on the cognitive normalcy of the offender, which, in this case, was absent. Rehabilitation was preferred to incapacitation as the method of protecting the public from the respondent because he was a young offender. In this regard, the Prosecution was wrong to emphasise that the court was fundamentally engaged in a “risk assessment” exercise because this failed to engage the specific principles justifying the importance of rehabilitating young offenders: at [115] and [121]–[124].
-
Retribution did not easily lend itself to being treated as a dominant sentencing objective because it simply required, in the context of sentencing, that the sentence imposed be commensurate with the offender’s culpability and the harm caused. But it was possible for retribution to be “displaced” as a sentencing consideration to some degree where the court was satisfied that the punishment should not correspond completely to the offender’s culpability and the harm caused, in order, for example, to achieve an outcome-focused objective such as incapacitation. In such cases, it was especially important that the court carefully apply any applicable sentencing framework and the totality principle to ensure that all things considered, the punishment fit the crime: at [131]–[133].
Issue 5: The appropriate sentencing option
-
The Court considered probation inappropriate in this case because it did not recognise the seriousness of the respondent’s offences, nor would it provide the respondent the structured environment that was necessary to manage his high risk of recidivism. Long-term imprisonment with caning was also inappropriate because it was principally a deterrent sentence and a form of incapacitation, and these objectives had no rational application to the respondent. While the Prosecution claimed that customised rehabilitation could be provided in prison, it did not explain why that could not similarly be provided in reformative training, given that both regimes were run by the Singapore Prisons Service: at [135]–[140].
-
The Prosecution’s proposed sentence was disproportionate. It was based on sentencing frameworks that were not promulgated with intellectually disabled offenders such as the respondent in mind. In addition, it violated both limbs of the totality principle, in that: (a) it was significantly higher than the usual range of sentences imposed for the most serious of the respondent’s offences; and (b) it was also crushing on the respondent, and not in keeping with his past record and future prospects. On the other hand, reformative training was proportionate, having regard to the sentences imposed for offences committed in similar circumstances. Furthermore, limitations in the sentencing regime were no justification for disproportionate sentencing: at [146]–[153], [157] and [158].
Conclusion
-
In the end, the Court was faced with a choice between imposing a term of between 15 and 18 years’ imprisonment and at least 15 strokes of the cane, and imposing a term of incarceration at a reformative training centre of up to three years. If those were both sub-optimal options, then in all the circumstances, the latter was the less imperfect and only principled option, especially considering that the respondent had already been incarcerated for almost four years. The sentence of reformative training could not and would not be backdated, and in all the circumstances, the Court was satisfied that this was the appropriate course: at [159].
This summary is provided to assist in the understanding of the Court’s grounds of decision. 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 grounds of decision.
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 3: 27 Oct 2020 (22:41 hrs)