1 These appeals concern the constitutionality of s 377A of the Penal Code (Cap 224, 2008 Rev Ed) (referred to hereafter as “s 377A” and “the PC” respectively), an issue which is before this court not for the first time. The appellants contend that s 377A is unconstitutional by reason of its inconsistency with Arts 9, 12 and/or 14 of the Constitution of the Republic of Singapore (1985 Rev Ed, 1999 Reprint) (“the Constitution”). Although s 377A is a law that, on its face, purports to do no more than prohibit particular sexual acts, its penalty and purpose carry more profound consequences, touching upon “the most private human conduct” and often in the context of the most private relationships (see Lawrence et al v Texas 539 US 558 (2003) at 567). Yet, for a law that has come to be intensely personal to many for what it means both practically and symbolically, it has also assumed an exceedingly public dimension. What rights the appellants claim to be fundamental, others view as controversial; what is to the appellants deeply personal and even definitive of their identity, others regard as offensive. On both sides of this divide, the continued existence of s 377A in our statute books has taken on particular importance because of what it is thought to signify. (For completeness, we note that the current s 377A of the Penal Code 1871 (2020 Rev Ed) (“the Revised PC”) is in pari materia with s 377A of the PC, which was the version of the provision that was in force at the time of the appellants’ respective applications and the subject of the present constitutional challenges. Likewise, the present Arts 9, 12 and 14 of the Constitution of the Republic of Singapore 1965 (2020 Rev Ed) (“the Revised Constitution”) are in pari materia with the corresponding Articles of the Constitution that were in force at the material time. Our judgment is thus equally applicable to the current s 377A of the Revised PC and Arts 9, 12 and 14 of the Revised Constitution.)

2 Before delving into the questions that these appeals present, we first clarify the scope of these appeals, beginning with what they are not about. They are not about whether s 377A should be retained or repealed, that being a matter beyond our remit. Nor are they about the moral worth of homosexual individuals. In the words of our Prime Minister, Mr Lee Hsien Loong (“the Prime Minister”), homosexual individuals are “part of our society” and “our kith and kin” (see Singapore Parliamentary Debates, Official Report (23 October 2007) vol 83 (“the 23 October 2007 Debates”) at col 2398 (Lee Hsien Loong, Prime Minister and Minister for Finance)). They are also not about the fundamental nature of sexual orientation (whether immutable or not), which is an extra-legal question well beyond the purview of the courts.

3 What, then, are these appeals about? The deceptively easy answer – namely, whether s 377A is inconsistent with the Constitution – belies the underlying complexity of the issues that are before us. Section 377A has long been a lightning rod for polarisation, in large part because it raises a wider question, which admits of no ready answers, of how a State can best maintain harmony between different communities with deeply held, and sometimes conflicting, views on important issues of moral conscience. As the socio-political debate over s 377A continues, the balance between various interests in society grows more delicate, and the need to create space for peaceful co-existence too becomes more pressing.

4 One may well ask whether litigation is, in fact, the optimal way to resolve such differences. Politics seems the more obvious choice than litigation for debating and resolving highly contentious societal issues. At the heart of politics lies the project of democratic engagement, as politicians aim to persuade voters by appealing to hearts and minds. Litigation, on the other hand, is “not a consultative or participatory process” but “an appeal to law” (see Jonathan Sumption, Trials of the State: Law and the Decline of Politics (Profile Books Ltd, 2019) (“Trials of the State”) at p 65). The single biggest advantage of the political process – in fact, its raison d’être – is its ability to accommodate divergent interests and opinions (see likewise Trials of the State at p 65). However sub-optimally some may think politics performs that function, the courts can never discharge that function simply because it is not their constitutional role to mediate such differences in society. And this is so for good reason, because litigation is a zero-sum, adversarial process with win-lose outcomes. The political process, in contrast, seeks to mediate – it strives for compromises and consensus in which no one side has to lose all.

5 The forum in which we resolve sincere disagreements over issues of moral conscience matters for at least three reasons. First, the political process is by nature more suited for the resolution of such issues. The courts deal with the retrospective adjudication of rights and liabilities arising out of past events, whereas politics aims to forge consensus amongst individuals with different (and often conflicting) interests in order to create policies to govern future conduct (see Sundaresh Menon, “Executive Power: Rethinking the Modalities of Control”, Annual Herbert L Bernstein Lecture in Comparative Law delivered at Duke University School of Law on 1 November 2018, 29 Duke J Comp & Int’l L 277 (2019) (“the Bernstein Lecture”) at 300). As such, it is Parliament, and not the courts, that is best placed to devise a pluralistic vision that accommodates divergent interests.

6 Second, even those who place a higher premium on outcomes than processes have every reason to be invested in having social controversies resolved by robust public debate rather than by litigation. One need only consider the outsized consequences of the decision of the Supreme Court of the United States in Roe v Wade 410 US 113 (1973) (“Roe v Wade”) to comprehend this point. In the 1970s, state legislatures in the United States seemed to be leaning towards the gradual liberalisation of abortion statutes. However, that process was abruptly disrupted by the Supreme Court’s decision in Roe v Wade in 1973, in which the court held that the Due Process Clause in the Fourteenth Amendment to the United States Constitution provided a “right to privacy” that protected a pregnant woman’s right to choose whether to have an abortion. It has been suggested that by short-circuiting the process of democratic, organic change, Roe v Wade “stimulated the mobilization of a right-to-life movement and an attendant reaction in Congress and state legislatures”, thereby deferring stable settlement of what has now become an intractable issue (see Ruth Bader Ginsburg, “Some Thoughts on Autonomy and Equality in Relation to Roe v. Wade” 63 N C L Rev 375 (1985) at 381). Nearly 50 years later, the ramifications of that decision continue to reverberate in the legal challenges that are still being brought against it, and in how prominently it continues to feature in the public examination of candidates for appointment to the Supreme Court.

7 Third, the courts risk a diminution in their legitimacy if they are perceived as having overstepped their boundaries. This risk may materialise if the courts wade into matters that call for resolution by discussion, consensus and debate, rather than by a judgment handed down by a court. It can only be injurious to public confidence in the courts if polycentric matters are not debated and resolved by the many but summarily adjudicated by the few. Accordingly, judicial restraint should be prioritised in the face of disagreements that stem from incommensurable conceptions of the good.

8 In the final analysis, certain issues call for continued discussion and open-ended resolutions, rather than win-lose outcomes recorded in a court judgment. Of course, some may disagree. The appellants point to the decision of the Supreme Court of India in Navtej Singh Johar & Ors v Union of India thr Secretary, Ministry of Law and Justice [2018] 10 SCC 1, which decriminalised same-sex intercourse, and contend that the time has come for us likewise to declare s 377A unconstitutional. However, one cannot look to developments abroad without first appreciating the exceptional situation in Singapore. The merits of retaining s 377A were subject to robust and lengthy debate in Parliament in 2007, culminating in a uniquely Singaporean resolution: a political compromise in which s 377A would be retained because it was thought to bear important symbolic weight for the conservative mainstream in Singapore. Exceptionally, this was on the basis that s 377A would not be proactively enforced, so as to accommodate our homosexual kith and kin. To our knowledge, no other country has struck a similar political compromise in respect of laws resembling s 377A.

9 The aforesaid political compromise was conceived with the express intention of accommodating divergent interests, avoiding polarisation and facilitating incremental change. Its purpose was to keep s 377A as a matter within the democratic space. There are consequences in removing issues of such profound public and moral significance from the realm of democratic decision. As Chief Justice John Roberts pithily put it in Obergefell v Hodges 576 US 644 (2015) at 710, “[c]losing debate tends to close minds”.

10 Importantly, the political compromise on s 377A that was forged in 2007 and that has been upheld since then has radically altered the complexion of this provision, in terms of both its constitutionality and its consequences for those whom it might affect. Any discussion on the constitutionality of s 377A therefore cannot ignore the terms, the purpose and the consequences of this political compromise.

11 As the appeals before us raise matters of searing socio-political controversy, we think it especially important for us, first, to characterise the role of the court correctly. It is incontrovertible that the doctrine of the separation of powers is part of the basic structure of the Westminster constitutional model that Singapore adopts. Constitutions based on the Westminster model incorporate this doctrine so as to diffuse state power amongst different organs of State (see Mohammad Faizal bin Sabtu v Public Prosecutor [2012] 4 SLR 947 (“Mohammad Faizal bin Sabtu”) at [11]–[12]). It follows from this doctrine that the court must refrain from trespassing onto what is properly the territory of Parliament.

12 However, as we observed in Saravanan Chandaram v Public Prosecutor and another matter [2020] 2 SLR 95 (“Saravanan”) at [154] and Wham Kwok Han Jolovan v Public Prosecutor [2021] 1 SLR 476 (“Jolovan Wham”) at [26], that is not to say that Parliament’s actions are presumptively constitutional. The doctrine of the separation of powers is premised on the notion that the Judiciary, the Executive and the Legislature are co-equal branches. It follows that the court will not defer to the other branches where the constitutionality or legality of a measure or statutory provision is challenged in legal proceedings. Indeed, any such deference would be contrary to the doctrine of the separation of powers and would portend the institutional irrelevance of the court. As each organ of State has its own role and space (see Jolovan Wham at [27]), the focus should be on whether a particular issue falls within the institutional sphere of a particular organ of State and whether any action taken within that sphere is constitutionally permissible. It goes without saying that in deciding whether a law or measure is constitutional, the court is concerned only with legal matters, and extra-legal considerations are beyond its purview.

13 Counsel for the Attorney-General (“the AG”), Ms Kristy Tan SC (“Ms Tan”), contends that whether s 377A should be repealed is a matter that falls within Parliament’s constitutional role. She acknowledges, however, that while the court cannot evaluate the policy merits of s 377A, it may determine the constitutionality of that provision. We agree. In our judgment, the legality of the political compromise that was struck in 2007 as regards s 377A similarly falls within the court’s institutional sphere. The fact that an impugned action is of a political nature or has socio-political significance does not, in and of itself, mean that the legality or constitutionality of that action cannot be assessed by the court – in fact, quite the contrary. The inherently political nature of the compromise reached on s 377A does not preclude the existence or application of legal standards against which its legality or constitutionality can be judged.

14 That said, it is equally important to note that the law is not and cannot be the continuation of politics by other means. The court is not a front-runner for social change or an architect of social policy, and the consequences may well be dire if it were otherwise (see [6] above). It hence bears reiterating that any judicial ruling in respect of s 377A ought to concern only its legality or constitutionality and cannot extend to the policy merits or socio-political desirability of retaining or repealing s 377A. After all, the role of the court is not to design policy but simply to “say what the law is” (see Marbury v Madison 5 US 137 (1803) at 177).

15 In summary, the doctrine of the separation of powers calls for each branch to respect the institutional space and legitimate prerogatives of the others. While this means that each branch should not intrude upon the institutional space that properly belongs to the other branches, it also means that each branch must be allowed to exercise fully and fairly the powers it has been allocated (see the Bernstein Lecture at 303). In that light, we turn to the facts of the present appeals.

16 We begin by setting out the relevant background to these appeals. Section 377A was enacted by the Legislative Council of the Straits Settlements (“the Legislative Council”) in 1938 by way of s 3 of the Penal Code (Amendment) Ordinance 1938 (SS Ord No 12 of 1938) (“the 1938 Penal Code (Amendment) Ordinance”), and provides as follows:

17 On 22 October 2007, a petition to repeal s 377A was presented to Parliament. Although no vote was taken on that petition, s 377A was debated extensively in Parliament during the second reading of the Penal Code (Amendment) Bill (Bill No 38/2007) (“the 2007 Penal Code (Amendment) Bill”): see Singapore Parliamentary Debates, Official Report (22 and 23 October 2007) vol 83. For ease of reference, we hereafter denote these debates collectively as “the s 377A Debates”. Of especial note is that while s 377A was retained, the Prime Minister made clear that this was on the terms that it would not be proactively enforced (see the 23 October 2007 Debates at col 2402). In his speech during the s 377A Debates, the Prime Minister emphasised the need to forge a consensus gradually and stated that the Government had decided to “keep the status quo on section 377A” (see the 23 October 2007 Debates at col 2405).

18 The issue of the constitutionality of s 377A first came before our courts in Tan Eng Hong v Attorney-General [2013] 4 SLR 1059, in which the High Court held that s 377A was constitutional and did not offend Arts 9 and 12 of the Constitution. That decision was upheld on appeal in Lim Meng Suang and another v Attorney-General and another appeal and another matter [2015] 1 SLR 26 (“Lim Meng Suang (CA)”). In the intervening years since our decision in Lim Meng Suang (CA), s 377A has remained a controversial and socially divisive issue, with calls both for and against its repeal intensifying in recent years.

19 Amidst the ongoing discussion on the retention or repeal of s 377A, Attorney-General Mr Lucien Wong SC (“AG Wong”), in his capacity as the Public Prosecutor (“the PP”), clarified in 2018 the prosecutorial policy in respect of offences falling under that provision. In a press release dated 2 October 2018 (“the 2018 AGC Press Release”), AG Wong acknowledged the Government’s position that “the Police will not proactively enforce [s 377A], for instance by conducting enforcement raids” (see para 3 of the 2018 AGC Press Release). He clarified, however, that the police would investigate reported offences under s 377A if the surrounding circumstances so warranted (for instance, if minors were involved). He concluded by stating (at para 6 of the 2018 AGC Press Release):

AG Wong reiterated this position in a letter published under his name in The Straits Times on 6 October 2018 (“the 2018 Straits Times article”). These statements contemplate the non-prosecution of a subset of acts that may be caught under s 377A, namely, sexual acts between two consenting adult men in private (“the Subset”).

20 Against those background facts, we provide an overview of the proceedings below.

21 The three appellants are homosexual men. The appellant in CA/CA 54/2020 (“CA 54”), Dr Tan Seng Kee (“Dr Tan”), is a medical doctor and LGBT (lesbian, gay, bisexual and transgender) activist. The appellant in CA/CA 55/2020 (“CA 55”), Mr Ong Ming Johnson (“Mr Ong”), is an international disc jockey, while the appellant in CA/CA 71/2020 (“CA 71”), Mr Choong Chee Hong (“Mr Choong”), is a former executive director of Oogachaga Counselling and Support.

22 The appellants each filed an application challenging the constitutionality of s 377A on the basis that it violates Arts 9, 12 and/or 14 of the Constitution (referred to hereafter as “Art 9”, “Art 12” and “Art 14” respectively). They sought declaratory relief and/or the voiding of s 377A to the extent of any such inconsistency with the Constitution.

23 In the proceedings below, counsel for Mr Choong, Mr Harpreet Singh Nehal SC (“Mr Singh”), focused on the proper interpretation of s 377A and the alleged unconstitutionality of that provision under Arts 12 and 14. His submissions on the proper interpretation of s 377A were twofold. First, he contended that s 377A, properly construed, criminalises only commercial sexual activity between men (meaning male prostitution) and not private consensual sex acts of a non-commercial nature. Second, he highlighted that when s 377A was enacted in 1938, penetrative sex acts were already criminalised as the offence of “carnal intercourse against the order of nature” under s 377 of the Penal Code (Cap 20, 1936 Rev Ed) (referred to hereafter as “s 377” and “the 1936 PC” respectively), which remained an operative provision until its repeal in 2007. He argued that s 377A was only intended to criminalise sex acts that were not already offences under s 377 – namely, non-penetrative sex acts.

24 As to the alleged unconstitutionality of s 377A, Mr Singh submitted that the provision fails the “reasonable classification” test and therefore violates Art 12 because there is no rational relation between its legislative object (of curbing male prostitution) and its differentia (which was asserted to be gay and bisexual men). He also proposed that the court adopt a proportionality-based test in place of the “reasonable classification” test for the purposes of assessing the constitutionality of a statutory provision under Art 12. Additionally, he argued that s 377A is inconsistent with Art 14 because it curtails the expression of love and closeness through consensual acts of sexual intimacy between men, and therefore unlawfully derogates from the constitutional right to freedom of expression.

25 Counsel for Mr Ong, Mr Eugene Singarajah Thuraisingam (“Mr Thuraisingam”), primarily argued that s 377A is inconsistent with Art 9(1) because it is absurd and arbitrary in criminalising individuals by reason of their sexual identity. In this regard, he sought to establish the immutability of sexual orientation by adducing affidavit evidence from various experts. According to Mr Thuraisingam, the expert evidence discloses a scientific consensus that sexual orientation is immutable and is not caused or influenced by socio-environmental factors. He further contended that s 377A violates Art 12 not only because there is no rational relation between its differentia (namely, male-male sex acts) and its legislative object (namely, the expression of disapprobation of either gross indecency or homosexuality in general), but also because it lacks an intelligible differentia to begin with.

26 Counsel for Dr Tan, Mr Ravi s/o Madasamy (“Mr Ravi”), broadly focused on the absurdity and arbitrariness of retaining s 377A in the light of the Government’s avowed position that s 377A would not be proactively enforced where conduct falling within the Subset (as defined at [19] above) is concerned. He also aligned himself with Mr Singh’s and Mr Thuraisingam’s submissions that s 377A falls afoul of Arts 9, 12 and 14.

27 The High Court judge (“the Judge”) found that s 377A did not violate Arts 9, 12 and/or 14. He therefore dismissed all three applications brought by the appellants: see Ong Ming Johnson v Attorney-General and other matters [2020] SGHC 63 (“the Judgment”).

28 The Judge first sought to interpret s 377A by applying the three-step framework for statutory interpretation set out in Tan Cheng Bock v Attorney-General [2017] 2 SLR 850 (“Tan Cheng Bock”). Under the first step of that framework (“the Tan Cheng Bock framework”), the Judge found that the ordinary meaning of the words “gross indecency with another male person” in s 377A was sufficiently wide to cover both penetrative and non-penetrative sexual activity between men. He also found that the language of s 377A did not connote any limitation to only activities involving male prostitution (see the Judgment at [94]).

29 Under the second and third steps of the Tan Cheng Bock framework, the Judge examined the extraneous material adduced by the appellants. The extraneous material comprised legislative material relating to the enactment of s 377A in 1938 that had been considered by this court in Lim Meng Suang (CA), as well as new extraneous material that was before the court for the first time. The Judge found that the new extraneous material was, strictly speaking, irrelevant for the purposes of statutory interpretation under the Tan Cheng Bock framework because it failed to meet the requisite standard of relevance and reliability (see the Judgment at [52]). Nonetheless, out of an abundance of caution, he proceeded to consider all the extraneous material that was before him.

30 The Judge found that while the problem of male prostitution had caused the British colonial administration much consternation, there was no mention of male prostitution in any of the relevant legislative material. It therefore could not be inferred that male prostitution was the only mischief that had necessitated the introduction of s 377A. Further, the fact that a more precise legislative solution was not crafted to specifically tackle the problem of male prostitution, even though this could easily have been done, suggested that s 377A was intended to be of broader application (see the Judgment at [112] and [137]). The new extraneous material did not change his view (see the Judgment at [113], [118], [145] and [146]).

31 As for the argument that s 377A was only intended to criminalise non-penetrative sex acts between men, the Judge acknowledged that prior to the enactment of s 377A, penetrative sex between men, whether in public or in private, was already criminalised under s 377 (see the Judgment at [107]). Moreover, s 23 of the Minor Offences Ordinance 1906 (SS Ord No 13 of 1906) (referred to hereafter as “s 23” and “the MOO” respectively) criminalised, among other things, indecent behaviour in public (see the Judgment at [103]). The gap in the criminal legislation prior to the introduction of s 377A in 1938 was hence that non-penetrative sexual activity between men in private was not criminalised under either s 377 or s 23 (see the Judgment at [108]).

32 However, the relevant legislative material showed that s 377A was not intended to deal only with non-penetrative sexual activity between men in private. Part of the legislative material referred only to the need to supplement s 377, whereas another part spoke only of the need to supplement s 23. The legislative material, read holistically, militated in favour of the conclusion that s 377A broadened the scope hitherto covered by s 377 by covering not only penetrative sex, but also other (less serious) acts of “gross indecency” between men (see the Judgment at [131]–[133]). The Judge was of the view that the new extraneous material did not undermine his conclusion, which was the very conclusion that we had arrived at in Lim Meng Suang (CA) at [134] (see the Judgment at [134] and [144]). This interpretation of s 377A (as covering both penetrative and non-penetrative sex acts) also accorded with the ordinary meaning of “gross indecency” as ascertained by the Judge under the first step of the Tan Cheng Bock framework (see the Judgment at [109]–[110]). As for Mr Singh’s argument that s 377A did not cover penetrative sex acts because s 377 already covered such conduct, the Judge observed that the PC contained numerous examples of overlapping offences (see the Judgment at [130]). He thus held that s 377A overlapped with both s 23 and s 377, and that there was no legislative intent to exclude any such overlap (see the Judgment at [146(c)]).

33 In rejecting the argument that s 377A was intended to deal only with non-penetrative sexual activity between men in private, the Judge also noted that s 377A was based on s 11 of the English Criminal Law Amendment Act 1885 (c 69) (referred to hereafter as “s 11 (UK)” and “the 1885 UK Act” respectively), which was invoked to prosecute offences involving sodomy (see the Judgment at [44], [49] and [97]). He observed that the use of s 11 (UK) was not confined to cases involving male prostitutes nor to cases involving non-penetrative sex acts (see the Judgment at [123]). The Judge ultimately concluded that the legislative purpose of s 377A was “to safeguard public morals generally” (see the Judgment at [146(d)]). Section 377A furthered this legislative purpose by demonstrating society’s moral disapproval of male homosexual sex acts and by enabling the enforcement and prosecution of all forms of gross indecency between men (see the Judgment at [146(d)], [181], [189], [191] and [298]).

34 The Judge noted that, as a matter of stare decisis, he remained bound by this court’s holdings in Lim Meng Suang (CA) as to the purpose and scope of s 377A. However, even if he were not so bound, and even after taking into account the new extraneous material put forth by the appellants, he found no reason to depart from those holdings (see the Judgment at [143], [144], [305] and [306]).

35 Turning to the constitutional challenge to s 377A under Art 12, the Judge held that s 377A had an intelligible differentia under the “reasonable classification” test as it targeted homosexual sex acts between men. This differentia could not be said to be unintelligible because distinctions were drawn between men and women in other areas of Singapore law (see the Judgment at [171]–[174] and [178]). There was also a rational relation between this differentia and the object sought to be achieved by s 377A, which was the safeguarding of public morals generally (see the Judgment at [146(d)], [179]–[181] and [189]).

36 The Judge rejected the appellants’ arguments that s 377A was both over-inclusive and under-inclusive (see the Judgment at [183]–[194]). He held that there was a complete coincidence between the differentia in s 377A (namely, acts of gross indecency between men) and the legislative object that s 377A sought to achieve (namely, “the criminalisation of male homosexual conduct to safeguard public morals generally and reflect societal morality”: see the Judgment at [189]). Consequently, he held that there was a rational relation between the differentia in and the object of s 377A and that s 377A satisfied the “reasonable classification” test (see the Judgment at [194]). The Judge was of the view that, in any event, he was bound by this court’s finding in Lim Meng Suang (CA) that s 377A is not contrary to Art 12 (see the Judgment at [164]).

37 The Judge also rejected the suggestion that the court ought to adopt a higher standard of scrutiny in the form of a proportionality-based test when assessing the constitutionality of a statutory provision under Art 12. His primary reason was that such a test would necessarily involve reviewing the legitimacy of the object of the statutory provision in question, which would entail the risk of the court acting like a “mini-legislature” (see the Judgment at [216]).

38 In respect of the constitutional challenge to s 377A under Art 14(1)(a), the Judge applied the Tan Cheng Bock framework to interpret the meaning and scope of the phrase “freedom of … expression” as used in that Article. He acknowledged that the plain meaning of the term “expression” did not rule out the possibility of sexual intercourse being a form of expression (see the Judgment at [244]). However, he noted that the marginal note to Art 14, which formed part of the relevant context in interpreting Art 14(1)(a), made no mention of freedom of expression as a free-standing right (see the Judgment at [245]–[246]). This suggested to him that the right to freedom of expression was “something relating to or falling within the right to freedom of speech ie the verbal communication of an idea, opinion or belief” (see the Judgment at [246]). Applying the ejusdem generis principle of statutory interpretation, the Judge found that the ordinary meaning of the term “expression”, when read together with the term “speech”, “necessarily point[ed] towards some form of verbal communication” (see the Judgment at [247] and [249]). He therefore held that the right to freedom of expression could not be divorced from the right to freedom of speech (see the Judgment at [249]).

39 Although the appellants contended that such a reading would render the term “expression” in Art 14(1)(a) otiose, the Judge considered that there ought to be some allowance for surplusage or redundancy, given the “long history and borrowed phraseology” of Art 14(1)(a) (see the Judgment at [252]–[255]). His view was reinforced by the Report of the Constitutional Commission (27 August 1966) (Chairman: Wee Chong Jin CJ) (“the 1966 Constitutional Commission Report”), which omitted any mention of an independent, free-standing right to freedom of expression when considering the predecessor to Art 14(1)(a) (see the Judgment at [258]–[259]).

40 As for the appellants’ argument that sexual acts between consenting adult men conveyed meaning (such as love and affection) and were thus a protected form of expression, the Judge accepted that acts of physical intimacy might, in certain circumstances, serve as the means through which meaning was conveyed. He emphasised, however, that the right to freedom of expression under Art 14(1)(a) had to be understood in its proper context – namely, that it was encompassed within the right to freedom of speech. He also noted that an expansive reading of the right to freedom of expression could lead to absurd outcomes (see the Judgment at [261]–[263]).

41 Having found that the term “expression” was encompassed within the term “speech”, which he took to mean “the verbal communication of an idea, opinion or belief”, the Judge held that Art 14(1)(a) did not afford a constitutional right to engage in male homosexual sex acts as a form of “expression” (see the Judgment at [246] and [265]).

42 In respect of the constitutional challenge to s 377A under Art 9, the Judge was unconvinced that the scientific evidence before him pointed to any definitive conclusion on the immutability of sexual orientation. Instead, he found that the changeability and causes of sexual orientation remained highly controversial and were extra-legal arguments that did not come within the proper purview of the courts (see the Judgment at [273], [277] and [279]).

43 The Judge also rejected Mr Thuraisingam’s contention that s 377A criminalised a person on the basis of his sexual identity. Apart from the fact that there was no conclusive scientific evidence to show that homosexuality was immutable and/or solely caused by biological factors, an individual’s sexual identity was simply not an element of the offence under s 377A. Sexual orientation in and of itself was completely irrelevant under s 377A: a heterosexual man could equally be prosecuted under that provision (see the Judgment at [281]–[282]). The Judge was further of the view that Mr Thuraisingam were effectively advocating the recognition of an unqualified constitutional right to “personal liberty” on the basis of a person’s homosexual identity. He rejected that submission because unenumerated rights could not be specifically protected, and because many constitutional rights were qualified and not absolute (see the Judgment at [283]).

44 The Judge went on to consider the appellants’ arguments pertaining to the AG’s stated policy that s 377A would not be proactively enforced. In his view, the appellants’ real objection concerned that policy of non-enforcement rather than the inherent constitutionality of s 377A. These were, according to the Judge, separate and distinct issues: the manner in which a provision was enforced, even if arbitrary, could not, without more, result in the provision itself being rendered unconstitutional (see the Judgment at [287]). In any case, the AG had provided some guidance (through the 2018 AGC Press Release) on when the police would investigate possible offences under s 377A (see the Judgment at [288]). The Judge further held that the non-enforcement of s 377A was not incompatible with ss 424 and 17 of the Criminal Procedure Code (Cap 68, 2012 Rev Ed) (“the CPC”). He reasoned that given the AG’s avowed stance that it would not ordinarily be in the public interest to prosecute sexual activity between consenting adult men in private, there was no real risk of anyone being held liable for failing to report such activity, even though s 424 of the CPC imposed an obligation to do so (see the Judgment at [285] and [291]). In the same vein, even though the police were required under s 17(1) of the CPC to investigate suspected offences under s 377A, they retained the discretion under s 17(2) of the CPC to not do so (see the Judgment at [292]).

45 For the foregoing reasons, the Judge held that s 377A was constitutional and dismissed all three applications brought by the appellants.

46 The heart of Mr Singh’s case on appeal lies in the proper interpretation of s 377A. Mr Singh maintains that s 377A was only intended to cover non-penetrative sex acts between men. According to him, the extraneous material as well as the architecture of and the language used in the PC show that the offences under ss 377A and 377 were intended to be strictly non-overlapping. He also reiterates that the true mischief targeted by s 377A was male prostitution.

47 Additionally, Mr Singh submits that s 377A violates Art 12 on several levels:

(a) First, both men and women can commit acts of gross indecency. However, s 377A does not criminalise female-female acts of gross indecency, whether in private or in public, or, for that matter, male-female acts of gross indecency. Section 377A is thus severely under-inclusive and falls afoul of Art 12.

(b) Second, s 377A is also over-inclusive because it criminalises private sex acts in the interest of safeguarding public morality. The AG’s and the Government’s general stance of not enforcing s 377A in respect of private sex acts shows that the criminalisation of such acts does not further the object of safeguarding public morality generally.

(c) Third, the Judge erred in finding that there was a complete coincidence between the differentia in and the legislative object of s 377A. According to Mr Singh, the Judge’s finding is premised on characterising the legislative object of s 377A in a manner which includes the very differentia of that provision. He contends that framing the legislative object of s 377A in such a circular manner effectively denudes Art 12 of any real force.

At the hearing before us, Mr Singh and his co-counsel, Mr Jordan Tan (“Mr Tan”), focused on the arguments summarised at [46], [47(a)] and [47(c)] above.

48 It will be recalled that, in the proceedings below, Mr Singh propounded the adoption of a proportionality-based test in place of the “reasonable classification” test for the purposes of determining whether a statutory provision violates Art 12 (see [24] above). On appeal, he initially advocated the adoption of an “intermediate scrutiny” test to evaluate whether a statutory provision that discriminates on the basis of sex (such as s 377A) passes muster under Art 12. This test requires that the impugned sex-based differentia serve an “important state interest” [emphasis in original omitted] and be “substantially related” [emphasis in original omitted] to serving that interest. However, Mr Singh subsequently revised his position. In an aide memoire tendered to us at the hearing, he urged us to adopt a test of “substantial connection” instead. Under this test, where a statutory provision discriminates on the basis of sex, there must be a “substantial connection”, and not merely a “rational relation”, between that differentia and the legislative purpose of the provision. The “intermediate scrutiny” test and the “substantial connection” test are similar, save that the latter does not require that the impugned sex-based differentia serve an important state interest.

49 Mr Singh adds that the Judge erred in interpreting the term “expression” in Art 14(1)(a) restrictively to mean “some form of verbal expression” [emphasis in original omitted]. He avers that acts of sexual intimacy clearly fall within the ambit of Art 14(1)(a) as they are “a fundamentally important form of personal expression”. Accordingly, he submits that s 377A impermissibly derogates from the constitutional right to freedom of expression under Art 14(1)(a).

50 In the final analysis, Mr Singh submits that s 377A must either be construed in a manner consistent with the Constitution or be declared unconstitutional and struck down.

51 Mr Thuraisingam submits that the expert evidence adduced in the proceedings below shows, on a balance of probabilities, that sexual orientation cannot be wilfully changed and is not influenced by socio-environmental factors. In view of this, he argues that s 377A exposes a class of persons to the risk of incarceration on account of their sexual identity while failing to advance any compelling object. He therefore contends that s 377A is “absurd” (per Yong Vui Kong v Public Prosecutor and another matter [2010] 3 SLR 489 (“Yong Vui Kong (MDP)”) at [16]) and, consequently, not “law” for the purposes of Art 9(1).

52 According to Mr Thuraisingam, s 377A violates Art 9(1) for two further reasons:

(a) First, s 377A is “arbitrary” (per Yong Vui Kong (MDP) at [16]) because there is no evidence of any rational, logical or coherent differentia between male and female homosexual intimacy. Mr Thuraisingam also adopts Mr Singh’s arguments that: (i) there is no rational relation between the differentia in and the object of s 377A; and (ii) the Judge erred in framing the object of s 377A as “the safeguarding of public morals through the criminalising of [male homosexual] conduct” [emphasis added] (see the Judgment at [191]).

(b) Second, s 377A is contrary to the rule of law. The 2018 AGC Press Release states that the police will not “proactively enforce” s 377A and that it will not ordinarily be in the public interest to prosecute consensual sexual conduct between adult men in private. It also states, however, that the police may investigate complaints of offences under s 377A. Hence, a man who engages in conduct amounting to an offence under s 377A cannot reasonably foresee whether he will be investigated and/or prosecuted for such conduct, which is inconsistent with the rule of law.

53 For the foregoing reasons, Mr Thuraisingam’s position is that s 377A cannot be interpreted consistently with the Constitution, nor can the purportedly unconstitutional portions of s 377A be severed to “save” the remaining portions. Consequently, s 377A must be voided in its entirety.

54 Mr Ravi’s position is that s 377A should be declared void. As a general observation, Mr Ravi places greater weight on the legislative intention underpinning the retention of s 377A in 2007 than on the legislative purpose behind its enactment in 1938. While he acknowledged that the relevant parliamentary intention is usually to be discerned as at the time the statutory provision in question was enacted, he submitted during the hearing that the legislative purpose of a statutory provision could change over time and that any new legislative purpose could be taken into account when interpreting that provision.

55 Although Mr Ravi argued in his written submissions that s 377A violates Arts 9, 12 and 14, he focused on three main points in his oral submissions. First, he submitted that the differentia in s 377A is necessarily unintelligible because the s 377A Debates evince an express legislative intention for s 377A to be “legally untidy and ambiguous”. Second, he argued that the s 377A Debates suggest that the legislative purpose underlying the retention of s 377A was to discourage gay rights advocacy while accommodating homosexual individuals in society. Assuming that was indeed the relevant legislative purpose, there is no rational relation between that purpose and the differentia in s 377A (namely, male homosexual acts). Third, he contended that the AG’s and the Government’s stance of not enforcing s 377A has generated considerable legal uncertainty. According to Mr Ravi, not only is there ambiguity as to the interplay between s 377A and s 424 of the CPC, there is also uncertainty as to whether public servants and laypersons are legally bound under ss 119 and 176 of the PC respectively to report the commission of a s 377A offence. He submitted that such arbitrariness and uncertainty go towards the very constitutionality of s 377A and are not merely issues of administrative law.

56 Counsel for the AG, Ms Tan, makes five key arguments as regards the constitutionality of s 377A.

57 First, Ms Tan contends that the text and the context of s 377A, as well as the relevant extraneous material, show that s 377A covers both penetrative and non-penetrative sex acts and is not limited to sexual activity involving male prostitution.

58 Second, and relatedly, Ms Tan argues that s 377A satisfies the “reasonable classification” test. In this regard, she submits that the legislative purpose of s 377A should be discerned only at the time of its enactment in 1938. She emphasised, especially in her oral submissions, that the legislative purpose of s 377A at that time was to express societal disapproval of male-male sex acts specifically; this legislative purpose could be discerned principally from the text of s 377A, which criminalises only male-male acts of gross indecency. Accordingly, s 377A satisfies the “reasonable classification” test and is consistent with Art 12.

59 Third, Ms Tan submits that the scientific evidence is inconclusive as to the causes and the alleged immutability of sexual orientation. She further contends that the court is not the proper forum to determine such extra-legal issues of scientific controversy.

60 Fourth, Ms Tan argues that s 377A does not violate Art 9. According to her, Dr Tan and Mr Ong, the appellants in CA 54 and CA 55 respectively, are inviting this court to read a substantive unenumerated right of sexual liberty, identity or privacy into the Constitution. Article 9, however, only protects procedural rights, and there is no room to read unenumerated rights into the Constitution. Ms Tan also submits that the very premise of Dr Tan’s and Mr Ong’s contention that s 377A is “absurd” or “arbitrary” is erroneous to begin with because s 377A does not criminalise sexual orientation in and of itself. She further stresses that the non-enforcement of s 377A can have no bearing on its constitutionality.

61 Finally, Ms Tan adopts the Judge’s finding that s 377A does not violate Art 14 because the acts criminalised under s 377A do not concern speech or expression within the meaning of Art 14(1)(a). She also submits that, in any event, Art 14(2)(a), which permits restrictions on the right to freedom of speech and expression in the interest of public morality (among other things), is squarely engaged by s 377A.

62 Based on the parties’ written submissions, the overarching question in these appeals is whether s 377A is inconsistent with Arts 9, 12 and/or 14 and, hence, unconstitutional to the extent of any such inconsistency. To this end, five main issues arise for our consideration:

(a) whether sexual orientation is immutable;

(b) the proper interpretation of s 377A;

(c) whether s 377A violates Art 9;

(d) whether s 377A violates Art 14; and

(e) whether s 377A violates Art 12.

63 It will be readily apparent that aside from Mr Ravi, counsel (and, for that matter, the Judge) did not place much weight on the s 377A Debates that took place in 2007. However, as we mentioned earlier at [10] above, the constitutionality of s 377A cannot be assessed shorn of the political compromise that was struck in 2007 and upheld in the ensuing years. Given that representations were made by the Government and, subsequently, by AG Wong to the effect that s 377A would not be proactively enforced, it would be entirely artificial to ignore that fact and to analyse the constitutionality of s 377A as if this provision were liable to be enforced in the same manner as any other provision of the PC. Significantly, the representations made by the Government and AG Wong would have a practical impact on the lives of homosexual men, which cannot be disregarded.

64 Hence, quite apart from whether s 377A is constitutionally valid in the abstract, the more relevant and perhaps more critical question is whether the political compromise on s 377A is legally relevant – and, if so, what its precise legal consequences are. We therefore address this issue first before considering the issues outlined at [62] above.

65 A political compromise on s 377A was struck in 2007, although, as we explain below, it may not have had immediate legal effect. Section 377A was retained in our statute books, but on the express basis that it would not be proactively enforced. The rationale behind this political compromise was explained in the speeches of the Prime Minister and the other Members of Parliament (“MPs”) during the s 377A Debates.

66 This political compromise may well have informed the prosecutorial policy on s 377A in the years that followed. The PP’s exercise of his prosecutorial discretion is, after all, informed by the public interest, and the Prime Minister’s ministerial statement is the pre-eminent source from which the prevailing public policy may be discerned (see UKM v Attorney-General [2019] 3 SLR 874 (“UKM”) at [141]). However, the aforesaid political compromise took on a new legal significance in 2018, when AG Wong articulated a general policy of not prosecuting s 377A offences. The political compromise forged in 2007 and AG Wong’s representations in 2018 (as mentioned at [8] and [19] above respectively) collectively form the political package surrounding s 377A today. While one might argue that prosecutorial discretion is to be exercised by the AG of the day (such that AG Wong’s representations would not fetter the PP’s prosecutorial discretion), AG Wong’s representations are of legal significance and we elaborate on this below.

67 In our judgment, it would be contrived and unrealistic to ignore the totality of the political package when assessing the legality of s 377A. This is because the political package has, in effect, significantly altered both the way in which s 377A practically affects the lives of homosexual men and what the provision means in Singapore today. As we made clear to the parties, in particular, to Mr Singh:

68 The parties (apart from Mr Thuraisingam) agreed at the hearing that the constitutionality of s 377A could not be divorced from the political package. Mr Tan noted that it was “not controversial” that the political package forms part of the relevant legal framework that informs our assessment of whether s 377A is constitutional. Mr Singh likewise acknowledged that “[t]he [c]ourt cannot take a blinkered approach and ignore what happened in 2007”. In the same vein, Ms Tan conceded that we ought to “look at the context in which the constitutionality of [s 377A] is being challenged”. Moreover, in analysing how the political package has affected the constitutionality of s 377A under Art 12, she effectively intimated that the political package could and did have a bearing on whether s 377A is constitutional. Although Mr Thuraisingam questioned the relevance of the political package to the present appeals, there was an air of unreality in the approach that he urged us to adopt, namely, to assess the constitutionality of s 377A in total disregard of the political package. As we have already mentioned, the political package has rendered s 377A radically different in meaning and significance from any other provision of the PC. The precise terms and the legal consequences of the political package thus merit careful consideration.

69 For the avoidance of doubt, this section of the judgment is not concerned with the purpose behind the enactment of s 377A in 1938, but rather, the equally important purpose behind its retention in 2007. This is because, in the uniquely Singaporean context of the political compromise on s 377A, what is referred to as “Parliament’s intention” where this provision is concerned is informed by the purpose behind not only its enactment, but also its retention. A proper understanding of the political package is therefore critical in ascertaining the true parliamentary intent behind the retention of s 377A.

70 On 22 October 2007, Nominated MP Mr Siew Kum Hong presented a parliamentary petition for the repeal of s 377A (“the Petition”). The Petition contended that s 377A was an unconstitutional derogation from the guarantee of equality and the equal protection of the law set out in Art 12(1). The then Leader of the House moved a motion to have the Petition referred to Parliament for consideration at the second reading of the 2007 Penal Code (Amendment) Bill, so that matters in the Petition could be “thoroughly and properly debated, discussed and decided by Parliament” (see Singapore Parliamentary Debates, Official Report (22 October 2007) vol 83 (“the 22 October 2007 Debates”) at col 2122).

71 By way of background, a comprehensive review of the Penal Code (Cap 224, 1985 Rev Ed) (“the 1985 PC”) had taken place prior to the s 377A Debates. Following extensive public consultations and deliberation amongst the Ministers, the Government decided not to repeal s 377A and “to leave things be” (see the 23 October 2007 Debates at col 2397). Consequently, s 377A was considered in the context of a petition rather than as part of a Bill; the MPs were not asked to vote on the Petition and did not do so. That said, s 377A was debated rigorously and at length during the s 377A Debates, arguably to the point of overshadowing Parliament’s key agenda then, which was the amendment of the 1985 PC.

72 As the head of the Government, the Prime Minister addressed Parliament during the s 377A Debates. He explained that the Government would maintain the status quo by retaining but not proactively enforcing s 377A. This was because the Government took the view that a delicate balance had to be struck between accepting homosexual individuals as part of our society and respecting the more traditional views of mainstream society (see also UKM at [204]–[205]).

73 Early in his speech, the Prime Minister put forth the Government’s views on homosexuality, which he regarded as being reflective of the views of most Singaporeans. He stated as follows (see the 23 October 2007 Debates at cols 2397–2398):

74 At the same time, the Prime Minister took pains to reiterate that the purpose of retaining s 377A was not to discriminate against homosexual men, who are “part of our society” and “our kith and kin” (see the 23 October 2007 Debates at col 2398; see also [2] above). Recognising that “[t]hey, too, must have a place in this society, and they, too, are entitled to their private lives” (see the 23 October 2007 Debates at col 2399), the Prime Minister explained how the Government ensured that homosexual men had a place in our society (see the 23 October 2007 Debates at col 2401):

75 The Prime Minister reiterated the Government’s stance of not proactively enforcing s 377A later in his speech as follows (see the 23 October 2007 Debates at col 2402):

76 The retention of s 377A while not proactively enforcing it was aimed at striking a “balance” between “uphold[ing] a stable society with traditional, heterosexual family values” and allowing “space for homosexuals to live their lives and contribute to the society” (see the 23 October 2007 Debates at cols 2399–2340). The Prime Minister cautioned that repealing s 377A might send “the wrong signal that our stance has changed, and the rules have shifted” (see the 23 October 2007 Debates at col 2402). Forcing the issue could foreclose any possibility of reaching “an agreement within Singapore society”, because “[p]eople who are presently willing to live and let live will get polarised and no views will change” (see the 23 October 2007 Debates at col 2405). The Prime Minister concluded his speech by stating: “we have … been right to adapt, to accommodate homosexuals in our society, but not to allow or encourage activists to champion gay rights as they do in the West” (see the 23 October 2007 Debates at col 2407).

77 Many of these points about the political compromise on s 377A were reiterated in the speeches of other MPs, including that of the then Senior Minister of State for Home Affairs, Assoc Prof Ho Peng Kee, who moved the 2007 Penal Code (Amendment) Bill to its second reading (see the 22 October 2007 Debates at cols 2175–2202).

78 Reading the s 377A Debates as a whole, it is apparent to us that the debates did not concern the purpose of enacting s 377A in 1938, the scope of s 377A at the time of its enactment, or its differentia. Nor was the Government’s position informed by any of those matters. Rather, it is abundantly clear that the Government was focused on reaching a socio-political compromise that would balance various competing interests and accommodate differing perspectives on homosexuality. The political compromise that was eventually put forth by the Government and endorsed by Parliament (as summarised at [65] above) was a conscious and considered decision made after intense deliberation (see the 23 October 2007 Debates at cols 2397 and 2401–2405).

79 In deciding to “leave things be” (see the 23 October 2007 Debates at col 2397), it did not escape the Government’s attention that doing so would generate “legal untidiness and ambiguity” (see the 23 October 2007 Debates at col 2405). The Prime Minister addressed this directly by explaining that maintaining the status quo, while unsatisfactory, was preferable to risking societal division over s 377A. In his words, “[i]t is better to accept the legal untidiness and ambiguity. It works, do not disturb it” (see likewise the 23 October 2007 Debates at col 2405).

80 We have laid out the Prime Minister’s speech in some detail because it is legally significant in two respects. First, it is an express articulation of the public policy and the public interest in respect of s 377A – namely, the need to balance conflicting views on homosexuality in Singapore (see [76] and [78] above). As was held in UKM (at [137]), the three main criteria for assessing whether any material bears out an alleged public policy are authority, clarity and relevance. In our view, the Prime Minister’s speech fulfils all three criteria. The public policy espoused in the Prime Minister’s speech is relevant because it pertains specifically to s 377A. It is also clear and emanates from a constitutionally authoritative source – the Prime Minister’s parliamentary speech, delivered in his official capacity as the head of the Government (see UKM at [138], [141] and [142]).

81 Second, the Prime Minister’s speech set out the prevailing position in terms that homosexual men were not being and would not be harassed because s 377A would not be proactively enforced (see [74] above). We note, however, that the Prime Minister did not specify what he meant in saying that s 377A would not be “proactively” enforced. This is a point that we return to later at [99] below.

82 While the Prime Minister’s speech during the s 377A Debates is undoubtedly important, it must be borne in mind that ministerial statements about the general non-enforcement of s 377A, even those by the Prime Minister himself, do not have the force of law and do not bind the PP, who exercises his prosecutorial discretion independently (see Tan Eng Hong v Attorney-General [2012] 4 SLR 476 (“Tan Eng Hong (Standing)”) at [181]). In practice, however, we expect that the PP of the day, in exercising his prosecutorial discretion when dealing with s 377A offences, would have taken into account the Prime Minister’s speech, which is an authoritative expression of the public policy and public interest on this matter.

83 A major development in this narrative, which significantly affected the political package, was AG Wong’s issuance of the 2018 AGC Press Release and the 2018 Straits Times article. Significantly, they post-date this court’s decision in Lim Meng Suang (CA) and are being considered by this court for the first time.

84 On 2 October 2018, AG Wong issued the 2018 AGC Press Release. We reproduce its most salient portions below:

85 AG Wong emphasised that the PP exercises his independent and constitutionally protected discretion when deciding whether to prosecute a case involving an offence under s 377A. He concluded by providing the following guideline on how the PP would exercise his prosecutorial discretion in this regard:

86 The 2018 AGC Press Release was swiftly followed by the 2018 Straits Times article four days later on 6 October 2018. There, AG Wong expressed the Government’s and the PP’s “respective longstanding approaches to Section 377A cases” [emphasis added] in the following terms:

(a) First, “the Government’s position that the police will not proactively enforce Section 377A with respect to private acts had been made public since at least 2006”.

(b) In addition, the PP had “consistently taken the position that, absent other factors, prosecution under Section 377A would not be in the public interest where the conduct was between two consenting adults in a private place” [emphasis added]. AG Wong reiterated that this prosecutorial policy had existed “when Mr [V K] Rajah [SC] was the PP and remains so today”. Before us, Ms Tan confirmed that the PP still maintains the general position of not proactively enforcing s 377A and that this position is broadly in line with what was set out in Parliament in 2007.

87 The 2018 Straits Times article also referred to a 2008 case which AG Wong had mentioned in the 2018 AGC Press Release, and which involved an offender who had been charged under s 377A for performing oral sex on a minor in a public place. AG Wong highlighted that that case “illustrate[d] the Government’s and the [PP’s] respective longstanding approaches to Section 377A cases” and showed that these approaches were aligned.

88 In our judgment, AG Wong’s statements are legally significant in three respects. First, they contain guidelines on the exercise of prosecutorial discretion in relation to s 377A offences, which is a matter within the PP’s purview. These guidelines, in substance, constitute representations that s 377A will generally not be enforced in cases of sexual conduct between consenting adult men in private (that is to say, acts falling within the Subset as defined at [19] above).

89 Second, these guidelines and the policy stance they embody are broadly aligned with the public policy and public interest expressed by the Prime Minister during the s 377A Debates. AG Wong was, however, specifically concerned with the enforcement of s 377A by way of prosecution. In that sense, he appears to have been addressing an issue distinct from, although related to, enforcement in the sense of police investigations (see [74] above). This much is evident from para 6 of the 2018 AGC Press Release, where AG Wong stated that “[t]he Police’s exercise of its enforcement or investigative powers should … not be conflated or confused with the PP’s exercise of discretion to commence prosecution”.

90 Third, AG Wong stated that the PP’s position was consistent with the Government’s stance that s 377A would not be proactively enforced. Paragraph 5 of the 2018 AGC Press Release explicitly states that the fundamental principles pertaining to the PP’s exercise of his prosecutorial discretion in relation to s 377A offences “have been repeatedly affirmed by past and present Attorneys-General and have also been recognised and respected by the Government and Parliament”. This seeming alignment of positions was later reiterated in the 2018 Straits Times article. We add that the PP’s position, as enunciated by AG Wong, is aligned with not only that of the Government but also that of Parliament, given that Parliament chose to retain s 377A in 2007.

91 To summarise, it is evident from the political backdrop to the retention of s 377A that the constitutionality of the provision cannot be assessed in a vacuum, a point which both Mr Singh (as well as his co-counsel, Mr Tan) and Ms Tan conceded. In our judgment, the political reality surrounding s 377A can be encapsulated in three points.

92 First, although s 377A was retained in our statute books, this was on the terms that it would not be proactively enforced. The Government’s evident unwillingness to repeal s 377A signals its assessment that society has yet to adequately integrate the opposing views of mainstream conservatives and the homosexual community, as well as its awareness that our multi-racial, multi-lingual and multi-religious community remains vulnerable along such fault lines. The Government was especially cognisant that forcing the issue would polarise those who are “presently willing to live and let live” (see the 23 October 2007 Debates at col 2405; see also [76] above).

93 Second, the purpose behind the retention of s 377A in 2007 says nothing about the reason(s) for its enactment in 1938. Instead, the retention of s 377A was directed at addressing a deeply divisive socio-political issue in a pragmatic way. The decision not to repeal s 377A was a legislative one that was informed not by the purpose behind the enactment of the provision some seven decades earlier, but by the Government’s objective of striking an optimal compromise between competing interests in our society and accommodating differing perspectives on homosexuality (see [78] above).

94 Third, and flowing from our previous point, the purpose of the political compromise on s 377A that was reached in 2007 and elaborated on by AG Wong in 2018 was to strike a careful balance between the opposing interests of various groups. The retention of s 377A served to accommodate the views of the more conservative segments of society, while the caveat that s 377A would not be proactively enforced served to accommodate the interests of homosexual individuals and to allow them to live their lives in as full a space as is presently possible (see [76] above). The Government was clearly seeking to calibrate the pace at which the complex issue of whether s 377A should be retained or repealed was resolved. In this regard, a measured pace with a degree of accommodation was seen as the optimal course for Singapore.

95 Returning to the point we made at [67] above, we reiterate that it would be artificial and unrealistic to ignore the profound implications of these considered legislative and executive actions when assessing the legality of s 377A. It is also common ground between most of the parties that the representations made by AG Wong in 2018 cannot be said to be devoid of any legal force or effect. Ms Tan acknowledged at various points during the hearing that if the PP were to institute a prosecution under s 377A today, the affected individual may well choose to bring an administrative law challenge premised on something akin to a substantive legitimate expectation. Similarly, Mr Singh did not suggest that AG Wong’s representations were of no legal effect; he in fact implicitly conceded that the need to give “practical effect and permanence” to these representations was of utmost relevance. Instead, his real contention was that these representations were insufficient to insulate homosexual men from the threat of prosecution.

96 Although a key tenet of the political package was that s 377A would not be “proactively” enforced, there remains some uncertainty as to whether homosexual men may be liable to prosecution under s 377A, and if so, when they may be so liable. In broad terms, the contours of the political package give rise to three important issues.

97 First, any representations operate in the present and do not bind future Governments or AGs, or even the current Government and AG Wong. The Executive’s discretion to determine policy remains unfettered, and it can change its policy in respect of the enforcement of s 377A. Similarly, AG Wong or a future AG may choose to exercise the prosecutorial discretion in relation to s 377A offences differently, given that Art 35(8) of the Constitution vests the AG with “power, exercisable at his discretion, to institute, conduct or discontinue any proceedings for any offence” [emphasis added]. Indeed, it has been made clear by the AG himself that “no binding assurance could be given that no future prosecutions would ever be brought under s 377A” [emphasis in original] (see Tan Eng Hong (Standing) at [181]–[182]). This gives rise to the possibility that homosexual men may face the threat of prosecution under s 377A in the future.

98 This point was advanced in Mr Choong’s affidavit and raised by both Mr Singh and Mr Thuraisingam in their oral submissions. It was also accepted by Ms Tan, whose response was simply that any such retraction of the AG’s policy of not prosecuting s 377A offences generally might be grounds for a fresh constitutional challenge. If AG Wong or a future AG were to resile from the prosecutorial policy on s 377A that is currently in place to give effect to the political compromise that was forged in 2007, or if the Government were to change the stance encapsulated in that political compromise, the space presently accorded to homosexual men to live freely without interference may be eroded. In other words, the representations made by the Government and AG Wong as to the general non-enforcement of s 377A are only of practical effect until their maker chooses not to abide by them, a precarity which has been likened to “[h]aving a gun put to your head and not pulling the trigger” (see the 23 October 2007 Debates at col 2405).

99 Second, there is some uncertainty as to both the scope and the source of any legal protection from prosecution under s 377A. During the s 377A Debates, the Prime Minister made clear that the Government does not “proactively enforce” s 377A. This would suggest to a layperson that s 377A will generally not be enforced. However, while the AG’s general stance of non-enforcement (in the sense of non-prosecution) broadly dovetails with that of the Government (see above at [89]), there might be some gaps between their respective positions. This is unsurprising: the Prime Minister was making a policy statement in Parliament on behalf of the Government, whereas AG Wong was effectively providing the general public with guidelines on the exercise of the prosecutorial discretion in relation to s 377A offences. AG Wong was therefore rather more specific in stating that only a certain category of acts caught by s 377A, namely, sexual conduct between consenting adult men in private, would generally not be prosecuted under that provision. The Prime Minister’s speech understandably did not go into such detail.

100 On top of the ambiguity as to the scope of legal protection from prosecution under s 377A, the source of any such legal protection is also murky. We first note that the expressed intention of Parliament, as well as the Government’s position as set out in the Prime Minister’s speech during the s 377A Debates, are clearly matters which the PP will take into account in exercising his prosecutorial discretion in cases involving s 377A offences (see para 2 of the 2018 AGC Press Release). As we observed at [80] above, the Prime Minister’s speech in Parliament is an express articulation of the public policy and the public interest in this politically charged area of law. Since the PP is guided by the public interest when exercising his prosecutorial discretion, it is unsurprising that AG Wong’s representations in 2018 are broadly aligned with the position set out in the Prime Minister’s speech. However, the source of any legal protection from prosecution under s 377A is uncertain, given that statements by the Government do not bind the PP’s exercise of his prosecutorial discretion. In this regard, it is pertinent that AG Wong stated in no uncertain terms that “[w]hile the PP is entitled to consider public policies in exercising his discretion, these do not fetter the exercise of prosecutorial discretion” [emphasis added] (see para 4 of the 2018 AGC Press Release).

101 AG Wong’s position is, of course, legally sound. As we mentioned above, the PP’s exercise of his prosecutorial discretion is protected under Art 35(8) of the Constitution. This prosecutorial function is completely insulated from all other parts of the Executive and is constitutionally vested solely in the AG (in his capacity as the PP), leaving him with full power to decide on all matters concerning the institution, conduct and termination of prosecutions. It is a matter of settled law that the PP makes all prosecutorial decisions without interference from other parts of the Executive and that his decisions can only be challenged on limited grounds such as unconstitutionality and abuse of prosecutorial power (see Ramalingam Ravinthran v Attorney-General [2012] 2 SLR 49 (“Ramalingam”) at [17], citing Law Society of Singapore v Tan Guat Neo Phyllis [2008] 2 SLR(R) 239 at [149]).

102 In our judgment, because of the AG’s exclusive constitutional responsibility for prosecutions (in his capacity as the PP), it is his position, and not the Government’s or Parliament’s, which is material for our purposes. Absent AG Wong’s express representations in 2018, which cohered with the political compromise on s 377A that was reached in 2007, it would not have been legally possible to limit the PP’s prosecutorial discretion in relation to s 377A offences by reference to that political compromise only. As the PP acts independently of the other parts of the Executive when exercising his prosecutorial discretion, he is not bound by the statements of Ministers, even those of the Prime Minister, as to how the Government or its officials will act (see Tan Hon Leong Eddie v Attorney-General [2021] SGHC 196 at [29]). Subject to the limited grounds of judicial review laid out in Ramalingam at [17], only the AG (in his capacity as the PP) can decide how his office will act and, more significantly, craft a policy regarding how his office will act in certain circumstances. In examining the legal consequences of the political package, therefore, it is AG Wong’s representations in 2018 as to the general non-enforcement of s 377A that are material. According to AG Wong, those representations were intended to be aligned with the Government’s position as expressed by the Prime Minister during the s 377A Debates.

103 This brings us to the third issue, which concerns the ambiguity as to how the political package affects the enforcement of other laws, a point which Mr Ravi stressed in both his oral and his written submissions.

104 The Prime Minister himself noted in his speech that the political compromise pertaining to s 377A was legally untidy. Among other things, the political package as a whole leaves it unclear how other laws will be applied or enforced. This was foreshadowed during the s 377A Debates by the then MP for Bishan-Toa Payoh, Mr Hri Kumar Nair, who said: “… in the long run, making some conduct criminal under our Penal Code whilst stating that the law will not be enforced, simply invites attacks on the integrity of the Code” (see the 22 October 2007 Debates at cols 2287–2288). The difficulties presented by the political package can be illustrated with reference to three discrete statutory provisions: s 424 of the CPC and ss 119 and 176 of the PC.

105 Section 424 of the CPC provides as follows:

106 As for ss 119 and 176 of the PC, they state as follows:

107 There remains uncertainty as to whether individuals who are aware of the actual or intended commission of conduct amounting to an offence under s 377A are obliged to report this to the police under s 424 of the CPC, absent a reasonable excuse. It is similarly questionable whether a public servant (such as a police officer) who knows that such conduct is likely to be committed is legally bound under s 119 of the PC to disclose the existence of a design to commit such conduct. We note that the voluntary concealment of information relating to the existence of a design to commit an offence under s 377A may attract an imprisonment term of up to one year (this being half the maximum term of imprisonment provided for under s 377A). Further, unlike s 424 of the CPC, s 119 of the PC does not provide for a defence of reasonable excuse. It is also uncertain whether individuals are legally bound under s 176(2) of the PC to provide public servants with notice or information for the purpose of preventing the commission of a s 377A offence or apprehending a person who has committed such an offence. AG Wong did not expressly deal with these points when making his representations in 2018, but this is unsurprising given that these points were only ancillary to the thrust of his representations.

108 The question thus remains open as to whether, despite AG Wong’s representations, the legal duties under ss 119 and 176 of the PC and s 424 of the CPC continue to apply in relation to conduct that might amount to an offence under s 377A, or whether this would be so only where the conduct involves acts committed in public and/or with a minor. Despite AG Wong’s intention to set out a prosecutorial policy in relation to s 377A that is congruent with the Government’s policy stance and articulation of the public interest (as summarised at [80] and [92]–[94] above), some gaps remain, although there is nothing to suggest that these gaps were intentional or deliberate.

109 In these circumstances, the legal framework that is based on the political compromise forged in 2007, as elaborated on by AG Wong in 2018, leaves the legal status of s 377A uncertain in some ways. Such uncertainty leaves homosexual men unable to plan their lives adequately as they do not know, with reasonable certainty, how s 377A will be applied or enforced. This is undesirable from a rule of law perspective: after all, it is a fundamental tenet of the rule of law that the law must be capable of guiding the conduct of those that it binds (see Tom Bingham, The Rule of Law (Penguin Books, 2011) at pp 91–95). The need for certainty is heightened by the fact that s 377A not only concerns matters as intensely private as acts of sexual intimacy, but also carries the threat of substantial punishment (of imprisonment for a term of up to two years). Even if prosecutions are not instituted under s 377A, this still leaves open the question of possible liability under ss 119 and 176 of the PC and s 424 of the CPC.

110 These rule of law concerns do not, however, lead to the conclusion that the political package should be discarded. Rather, given the signal importance of the political compromise on s 377A that was struck in 2007, the court should strive to honour and give legal effect to that compromise as far as practicable. In our judgment, there are two key reasons why the political package should be upheld (with any unintended gaps filled, if necessary), rather than jettisoned.

111 First, the political package was clearly intended to assure homosexual men that they have a place in our society and that they would not be “harass[ed]” despite the retention of s 377A. It is difficult to reconcile the publication of AG Wong’s representations in 2018 with any other view. Discarding the political package would mean depriving homosexual men of the legally binding assurance that they can live freely in Singapore, without harassment or interference.

112 Second, Parliament made a deliberate and considered choice to retain s 377A as part of the political package. If the political package can be upheld so as to give legal effect to the Government’s avowed position and the PP’s prosecutorial policy on s 377A offences (which is broadly aligned with the Government’s stance), then the court should strive to do so, albeit with the necessary finetuning to make it legally workable. To choose instead to embark on an arid legal analysis driven by the attempt to establish the motivations of the colonial draftsmen, while ignoring the far more current debates and resolutions of our Parliament and our Government, would be to miss the forest for the trees. However, we are not thereby incorporating what is called the “living tree” doctrine (see Henrietta Muir Edwards and others v The Attorney-General of Canada and others [1930] AC 124), or any other doctrine akin to it, in the interpretation of our Constitution or statutes. Rather, the interpretation of s 377A as informed by Parliament’s intention to retain the provision is necessitated by the unique circumstances in which the political package was arrived at.

113 We emphasise again that we are concerned with the enforcement of s 377A only in the sense of prosecution and not in any other sense (such as, for example, the conduct of police investigations). This is because some acts that are captured by s 377A may well constitute other offences (see below at [137]). As the exact offence that has been committed will often not be apparent at the outset of investigations, law enforcement agencies cannot and should not be constrained from investigating any reports of suspected offences.

114 To consider how the political package may be given legal effect, it is helpful to reiterate its main features, which are as follows:

(a) Section 377A was not repealed in 2007 because Parliament judged that it was undesirable to do so then (and presumably continues to hold this view). Repealing s 377A might signal a change of the Legislature’s stance, which could in turn polarise the community and deepen social schisms, something that we can ill afford as a nation.

(b) The question of when it might be appropriate to repeal s 377A is self-evidently one of policy that the Legislature and the Executive are best placed to decide through the public forum of democratic discourse.

(c) At the same time, as an accommodation of our homosexual kith and kin, s 377A would not be proactively enforced by the Executive.

(d) The Government’s position in 2007 was animated by the desire to ensure that homosexual individuals would not be harassed and could live freely within the space afforded to them. What follows from this is the need to ensure that homosexual individuals do not face unreasonable legal risks that might be regarded as giving rise to a sense of harassment or curtailing their ability to live peacefully in our society.

(e) At the same time, the accommodation of homosexual individuals (by not proactively enforcing s 377A) was not intended to expand the space that they had been afforded. Any such expansion is a matter for Parliament to decide in the light of societal trends.

(f) While the Prime Minister left open the meaning of “proactive” enforcement, it seems possible to infer that private consensual sex acts between male adults (that is to say, acts falling within the Subset) would generally fall outside the ambit of enforcement. On the other hand, homosexual conduct in public or involving minors would not be insulated from enforcement. It should be noted that conduct amounting to offences under s 377A may also be prohibited under other laws.

115 It seems to us that these were the essential features of the political package that AG Wong’s representations in 2018 and the guidelines contained therein sought to give effect to.

116 As we have noted, the parties mostly agree that AG Wong’s representations must be of some legal effect and that the constitutionality of s 377A cannot be analysed in complete disregard of those representations. Therefore, before we consider the constitutionality of s 377A under Arts 9, 12 and 14, it is incumbent upon us to resolve the anterior question of the basis upon which AG Wong’s representations may be said to possess legal force and the precise legal effect and consequences of those representations (“the Anterior Question”).

117 In our judgment, the exceptional circumstances surrounding the general non-enforcement of s 377A call for a limited recognition of the doctrine of substantive legitimate expectations as the basis for imbuing AG Wong’s representations with legal force.

118 In the sections that follow, we first examine the treatment of this doctrine in Singapore before explaining why a limited recognition of this doctrine is both permissible and warranted in the specific context of s 377A.

119 The classic exposition of the doctrine of substantive legitimate expectations can be found in Regina v Inland Revenue Commissioners, Ex parte MFK Underwriting Agents Ltd [1990] 1 WLR 1545 at 1568–1569, where Bingham LJ (as he then was) stated:

This same passage was subsequently cited in the seminal decision of the United Kingdom Supreme Court in In the matter of an application by Geraldine Finucane for Judicial Review (Northern Ireland) [2019] UKSC 7 (“Finucane”) at [55].

120 In SGB Starkstrom Pte Ltd v Commissioner for Labour [2016] 3 SLR 598 (“Starkstrom”), we distilled the essence of the doctrine of substantive legitimate expectations in the following terms (at [41]):

121 The doctrine of substantive legitimate expectations has evolved over the years and has been adopted and adapted in various common law jurisdictions (see Finucane at [62] for the most recent restatement of the doctrine in the United Kingdom; see also the seminal decision of the Hong Kong Court of Final Appeal in Ng Siu Tung & Others v Director of Immigration (2002) 5 HKCFAR 1 at [91]–[98] and [101]–[104], which was recently followed in U Storage Group Ltd v Director of Fire Services [2020] HKCFI 2114 at [113] and [115]–[117]).

122 However, the doctrine has not been uniformly well-received. In Canada, legitimate expectations operate only as a facet of procedural fairness and cannot give rise to substantive rights (see Agraira v Canada (Public Safety and Emergency Preparedness) [2013] 2 SCR 559 at [97] and, more recently, Dabao v Investigation Committee of the Saskatchewan Registered Nurses’ Assn [2020] SKQB 242 at [33]). In Australia, the terminology of legitimate expectations has fallen out of favour, having been described by the High Court of Australia as “apt to mislead”, “unsatisfactory” and “superfluous and confusing” (see Minister for Immigration and Border Protection v WZARH and Another (2015) 326 ALR 1 (“WZARH”) at [28], citing South Australia v O’Shea (1987) 163 CLR 378 at 411 and 417 as well as Attorney-General (NSW) v Quin (1990) 170 CLR 1 at [55]). The High Court of Australia has gone so far as to say that even though administrative decision-makers must ensure procedural fairness to those affected by their decisions, “[t]he ‘legitimate expectation’ of a person … does not provide a basis for determining whether procedural fairness should be accorded to that person or for determining the content of such procedural fairness” (see WZARH at [30]). The court further warned (likewise at [30] of WZARH) that recourse to the doctrine is “both unnecessary and unhelpful” and “may well distract from the real question; namely, what is required in order to ensure that the decision is made fairly in the circumstances having regard to the legal framework within which the decision is to be made”. The decision in WZARH was recently followed in Bell v Native Title Registrar [2021] FCA 229 at [74].

123 This court has yet to rule definitively that the doctrine of substantive legitimate expectations is part of Singapore law. To explain this, it is necessary for us to refer to two cases in detail. The first is the High Court’s decision in Chiu Teng @ Kallang Pte Ltd v Singapore Land Authority [2014] 1 SLR 1047 (“Chiu Teng”), in which it was held for the first time that the doctrine applied in Singapore. The second is our decision in Starkstrom, where we highlighted the difficulties in accepting the doctrine as part of Singapore law and deferred resolution of the matter to a suitable occasion in the future.

124 In Chiu Teng, the applicant sought to invoke the doctrine of substantive legitimate expectations to compel the Singapore Land Authority (“SLA”) to act in accordance with its representations as to how it would calculate the differential premium payable for state leases. The SLA had published two circulars and maintained a website with information on how the applicable differential premiums would be calculated. On that basis, the applicant argued that the SLA should not be permitted to act in a manner contrary to the legitimate expectations that its representations had allegedly engendered.

125 Following a comprehensive review of the positions in Australia, Hong Kong, the United Kingdom and Canada, the High Court held (at [119]) that “the doctrine of legitimate expectation should be recognised in our law as a stand-alone head of judicial review and substantive relief should be granted under the doctrine subject to certain safeguards”. Those safeguards were enumerated as follows (likewise at [119]):

126 The High Court considered that it was “eminently within the powers of the judiciary” to uphold legitimate expectations (at [113]), and explained the normative reasons for accepting the doctrine of substantive legitimate expectations as part of Singapore law as follows (at [112]):

On the facts, however, the High Court held that there were no grounds for invoking the doctrine.

127 Two and a half years after the decision in Chiu Teng, we considered the doctrine of substantive legitimate expectations in Starkstrom. That case arose from a serious workplace accident suffered by one of the appellant’s employees, whose injuries rendered him mentally incapacitated. The employee’s brother brought a claim for statutory compensation under the Work Injury Compensation Act (Cap 354, 2009 Rev Ed) (“the WICA”) on the employee’s behalf in May 2010, even though he had yet to be appointed as the employee’s deputy then. The claim was accepted by the Commissioner for Labour (“the Commissioner”), who was the respondent in the proceedings, and a notice of assessment was issued pursuant to s 24(2) of the WICA. The employee’s brother subsequently sought to withdraw the claim. He contended that the claim was legally invalid because he had only been appointed as the employee’s deputy in August 2012, after the claim had been brought. The Commissioner initially maintained that the claim was valid but subsequently decided that it was invalid and that the notice of assessment had been issued in error. The appellant commenced judicial review proceedings to quash the Commissioner’s eventual decision that the notice of assessment was a nullity for having been issued in error. It argued that it had a substantive legitimate expectation that the claim brought by the employee’s brother was valid and ought to be upheld.

128 We concluded that the doctrine of substantive legitimate expectations was inapplicable on the facts of the case because the dispute was between two private individuals, rather than between a public authority and an individual (at [42]). Additionally, the Commissioner had not made any actionable representations that engaged the doctrine (at [46], [47] and [49]). In making these findings, we assumed, without deciding, that the doctrine was part of Singapore law (at [41]). This was because it was unnecessary for us to either affirm or overrule Chiu Teng as to the existence of the doctrine as part of Singapore law. Nonetheless, we made the following important observations:

(a) First, the court’s role in a judicial review application is a limited one that concerns “a review of the decision-making process, as opposed to the merits of the decision” [emphasis in original] (at [56]). This flows from the doctrine of the separation of powers, the need to uphold Parliament’s intention (as expressed in statute) to vest certain powers in the Executive and pragmatic concerns over institutional competence (at [58]).

(b) Second, recognising the doctrine of substantive legitimate expectations as part of Singapore law would represent “a significant departure from [the] current understanding of the scope and limits of judicial review”. That would potentially change the understanding of the courts’ role in undertaking judicial review of administrative or executive actions, and could cause the courts to redefine their approach to the doctrine of the separation of powers as well as the relative roles of the judicial and the executive branches (at [59]).

(c) Third, the doctrine has not been uniformly well-received across common law jurisdictions (at [60]). We have already noted this at [122] above.

(d) Fourth, if the doctrine were to result in the grant of additional protection to private individuals in relation to public authorities, it would entail more searching scrutiny of executive action, beyond the traditional heads of judicial review (namely, irrationality, illegality and procedural impropriety). This might result in the court having to weigh private interests against public interests, which would, in turn, raise questions regarding the separation of powers and institutional competence (at [62]).

129 We expressed the view that “the difficulties inherent in accepting the doctrine of [substantive] legitimate expectations in Singapore should not be underestimated” (at [60]), and outlined a list of thorny issues that “remain[ed] to be determined” should the issue of incorporating the doctrine into our law arise in the future (at [61]):

130 We then concluded by framing the central question arising from the issues listed in the above passage as follows (at [62]):

131 In the years following Starkstrom, our courts have continued to leave open the question whether the doctrine of substantive legitimate expectations is or should be part of Singapore law (see Kardachi, Jason Aleksander v Attorney-General [2020] 2 SLR 1190 at [56]–[57] and Tan Liang Joo John v Attorney-General [2020] 5 SLR 1314 at [61]–[62]).

132 In that light, we return to the expectations of homosexual men that s 377A will generally not be enforced in the context of acts falling within the Subset. The reasons why these expectations merit legal protection have been canvassed at [111]–[112] above and we do not propose to repeat them. Instead, we explain why a limited recognition of the doctrine of substantive legitimate expectations is appropriate in the unique circumstances surrounding the general non-enforcement of s 377A and serves as the basis for imbuing the representations made by AG Wong in 2018 with legal force. In doing so, we also provide our answer to the Anterior Question as framed at [116] above. To be clear, we recognise that AG Wong’s representations were made to a class of persons (namely, homosexual men) rather than to any particular individual. While some of the early jurisprudence on the doctrine of substantive legitimate expectations arose out of circumstances in which a public authority made certain representations to an individual (see [119] above), we consider that a representation made to a general class of persons may in principle give rise to a specific individual’s legitimate expectation, as long as that individual is uncontroversially a member of the class of persons to whom the representation was made.

133 Nevertheless, our recognition of the doctrine of substantive legitimate expectations is an extremely limited one and is shaped by two fundamental considerations. First, in the specific context of s 377A, a failure to recognise the legal effect of AG Wong’s representations may expose some individuals to the grave threat of prosecution and the attendant deprivation of liberty. In the light of the severe repercussions that might follow if AG Wong’s representations are not accorded legal effect, there is a strong impetus for recognising the doctrine of substantive legitimate expectations – albeit to a limited extent – in this specific context.

134 Second, and more importantly, the circumstances surrounding the general policy of not enforcing s 377A are exceptional. A decision was made in Parliament to strike a balance by preserving the legislative status quo on a vexed area of socio-political policy while accommodating the concerns of those directly affected by the legislation in question. This was done in order to avoid driving an irrevocable wedge within our diverse society. The first part of that balance, namely, the decision not to repeal s 377A, was effected by Parliament. What we are now concerned with is the second part of that balance, namely, the proper contouring of the accommodation that has been extended to homosexual men, which is well within our ambit to determine. This issue has not been judicially considered because, as we have explained, AG Wong’s representations post-date our decision in Lim Meng Suang (CA). In our judgment, by invoking the doctrine of substantive legitimate expectations in these unique circumstances, we would be upholding the public interest in maintaining the legislative status quo along the lines delineated by Parliament in 2007 and affirmed by AG Wong in 2018. Such an approach would also be consistent with avoiding a precipitous resolution of the issue of whether or not s 377A should be removed from our statute books, which Parliament itself had eschewed in 2007.

135 Recognising the doctrine of substantive legitimate expectations in these specific circumstances would, most atypically, neither offend the doctrine of the separation of powers nor require us to review the substantive merits of the political package. On the contrary, in giving effect to the expectations of homosexual men that s 377A will generally not be enforced in respect of acts falling within the Subset, the court would be giving effect to the political compromise on s 377A, as articulated during the s 377A Debates and reiterated by AG Wong in articulating the prosecutorial policy on s 377A. Critically, our holding would not constrain any future legislative or executive action regarding s 377A. In particular, there is no suggestion that the PP has acted or intends to act in a way contrary to the expectations that have been generated.

136 There is also no issue of a possible encroachment on the PP’s prosecutorial discretion. As the institution, conduct and termination of prosecutions lie within the exclusive domain of the AG (acting in his capacity as the PP), it would ordinarily be inappropriate for the court to restrict or interfere with that discretion by reference only to the Prime Minister’s parliamentary speech. Indeed, as was noted in Chiu Teng at [119(f)(i)], even if the doctrine of substantive legitimate expectations were part of Singapore law, it may not be invoked if doing so would result in a breach of the law. As it turns out, however, after our decision in Lim Meng Suang (CA), AG Wong published his office’s prosecutorial policy on s 377A, which was expressed to give effect to the public policy reflected in the Prime Minister’s speech during the s 377A Debates. Hence, and most exceptionally, giving effect to the legitimate expectations that have arisen as a result of the political package does not risk curtailing the PP’s prosecutorial discretion or offending the doctrine of the separation of powers. Furthermore, it certainly does not entail the court having to evaluate the merits of the positions adopted by AG Wong or, for that matter, by the Prime Minister.

137 We make two further points concerning the implications of our giving effect to AG Wong’s representations in the specific context of s 377A. First, s 377A covers a broad range of conduct, including conduct that is indisputably objectionable and deserving of legal sanction, such as the abuse of male minors or acts of gross indecency in public. AG Wong made clear in 2018 that it would not normally be in the public interest to institute prosecutions under s 377A in cases involving two consenting adult men who engage in sexual activity in private. That leaves no question that it would sometimes be in the public interest to prosecute conduct that might be caught by s 377A but does not involve sexual acts between consenting adult men in private. Such conduct includes, for example, sexual conduct involving minors and/or occurring in a public place. However, as much as such conduct is objectionable, it would also be caught under more targeted (and, by and large, gender-neutral) legislative provisions. For example, indecent behaviour in public is prohibited under s 20 of the Miscellaneous Offences (Public Order and Nuisance) Act 1906 (2020 Rev Ed) and s 294 of the Revised PC, while sexual acts involving minors are prohibited under provisions such as ss 376A to 376C of the Revised PC. Nothing in this judgment constrains the PP’s freedom and ability, even now, to institute prosecutions under other laws in respect of such objectionable conduct. Ms Tan accepted this at the hearing.

138 Second, there is nothing at present to suggest that AG Wong wishes to depart from the position that he articulated in 2018. It is clear, however, that AG Wong or a future AG cannot be prevented from changing that position in the future (see [97] above). While any potential change in the AG’s position is not an issue before us, it seems to us that, in such an event, what would be required as a matter of fairness is that the AG provides, in clear and unambiguous terms, reasonable notice of his intention to resile from the representations previously promulgated by AG Wong in 2018.

139 This follows from the notion that adequate notification of a relevant change in policy destroys any expectation founded upon an earlier policy (see H W R Wade & C F Forsyth, Administrative Law (Oxford University Press, 11th Ed, 2014) at p 454), as illustrated by the Privy Council’s decision in Fisher v Minister of Public Safety and Immigration and Others (No 2) [1999] 2 WLR 349. In that case, the appellant, who had been convicted of murder and sentenced to death, submitted a petition to the Inter-American Commission on Human Rights (“the IACHR”) to obtain its views on whether the carrying out of his death sentence would be a violation of his human rights. The Privy Council rejected the appellant’s argument that he had a legitimate expectation that he would not be executed pending the IACHR’s consideration of his petition. The Privy Council found (at 356E) that even if the appellant originally had such a legitimate expectation, that expectation could not survive the Bahamas Government’s letters of 2 and 30 January 1998 informing his solicitors that the Bahamas Government would wait until no later than 15 February 1998 for the IACHR’s decision on his petition.

140 Drawing together the points raised above, it is permissible to recognise the doctrine of substantive legitimate expectations in the specific context of s 377A. On the present facts, such a move will not engage the thorny issues raised in Starkstrom at [61]. In particular, we reiterate that our finding does not entail a substantive review of the merits of the political package or offend the doctrine of the separation of powers. In giving effect to the legitimate expectations arising from AG Wong’s representations, we are doing no more than giving effect to the political package that the other constitutional actors – namely, the AG and the Government – have crafted in tandem. Furthermore, the circumstances that obtain in respect of s 377A are unique for the reasons set out at [134]–[136] above. As these highly unusual circumstances will hardly, if ever, manifest themselves in other contexts, we are satisfied that the doctrine of substantive legitimate expectations should be recognised to the limited extent warranted by the present context. For the avoidance of doubt, we do not, by this judgment, import the doctrine into Singapore law in any wider context. We leave that matter open for consideration on a future occasion when it is necessary for us to make a determination and when we have had the full benefit of counsel’s submissions.

141 For the doctrine of substantive legitimate expectations to be successfully invoked, it is not enough that an expectation is found to exist – the expectation must also be legitimate and worthy of legal protection. In this regard, we agree broadly with the analytical framework that was set out in Chiu Teng at [119] (see [125] above). We have also outlined why we think the expectations of homosexual men that s 377A will generally not be enforced in respect of acts falling within the Subset are legitimate and deserve legal protection. However, a question remains as to whether an expectation merits legal protection only when it has been detrimentally relied upon.

142 In Chiu Teng, the High Court held (at [119(e)]) that an applicant had to establish detrimental reliance in order to invoke the doctrine of substantive legitimate expectations. The need to prove detrimental reliance has, however, been doubted in a series of cases in the United Kingdom (see Lord Kerr’s leading judgment in Finucane at [62], [70] and [72]; Lloyd Jones LJ’s opinion in Regina (Patel) v General Medical Council [2013] 1 WLR 2801 (“Patel”) at [84]; Lord Hoffmann’s opinion in R (on the application of Bancoult) v Secretary of State for Foreign and Commonwealth Affairs [2008] 4 All ER 1055 at [60]; Schiemann LJ’s opinion in Regina (Bibi) v Newham London Borough Council [2002] 1 WLR 237 at [31] and [55]; and Peter Gibson LJ’s opinion in Regina v Secretary of State for Education and Employment, Ex parte Begbie [2000] 1 WLR 1115 (“Begbie”) at 1124B–1124D).

143 Even if detrimental reliance is not an element of the doctrine of substantive legitimate expectations, it may be relevant in at least two ways, as stated in Begbie at 1124C–1124D. First, it may evidence the existence or extent of an expectation. Second, it may be relevant to a public authority’s decision whether to revoke a representation that forms the basis of an expectation. We add that detrimental reliance may also be an important consideration in determining the weight to be accorded to an expectation and, accordingly, where the balance of fairness lies (see Patel at [84]).

144 It is not necessary for us to come to a final view on whether detrimental reliance is an essential element of the doctrine of substantive legitimate expectations. Even if it is, there can be no question that homosexual men would have detrimentally relied on AG Wong’s representations – at least, by engaging in consensual homosexual sexual activity in private with the expectation that they would not be prosecuted for such conduct. As a general point, the reliance that individuals place on guidelines on the exercise of prosecutorial discretion, such as those that AG Wong’s representations constitute, cannot be understated. It is precisely because individuals detrimentally rely on such prosecutorial guidelines that in Regina (Purdy) v Director of Public Prosecutions (Society for the Protection of Unborn Children intervening) [2010] 1 AC 345, the House of Lords held (at 396) that the Director of Public Prosecutions had to provide a more detailed explanation of the factors that he would consider in deciding whether to prosecute offences of assisted suicide. Hence, regardless of whether detrimental reliance is an element of the doctrine of substantive legitimate expectations, we are satisfied that recourse may be had to the doctrine to the limited extent set out in this judgment.

145 We turn to examine the extent to which the doctrine of substantive legitimate expectations applies in this case. To begin, we highlight the following statements made by AG Wong in 2018:

(a) In para 6 of the 2018 AGC Press Release:

(b) In the 2018 Straits Times article:

146 In our judgment, the legitimate expectation engendered by AG Wong’s representations is that the PP will not prosecute conduct falling within the Subset under s 377A. We explain why with reference to the considerations set out in Chiu Teng at [119].

147 The import of AG Wong’s representations is that s 377A will not be enforced in respect of conduct falling within the Subset. There is no doubt that these representations were made by a person with actual authority to do so: they were made by AG Wong in his official capacity as the AG, in which capacity he exercised the powers and office of the PP. These representations were publicly promulgated and, pertinently, were made less than a month after the filing of the first of the three related originating summonses from which the present appeals stem. Furthermore, AG Wong’s representations are legally significant, not least because they constitute guidelines on how the PP will exercise his prosecutorial discretion in relation to s 377A offences. Indeed, and as we have already noted, the parties largely acknowledge the legal significance of these representations. Finally, giving effect to the legitimate expectation engendered by these representations, subject to the reservation noted at [138] above, will not result in a breach of the law or of Singapore’s international obligations. Nor is there any indication that giving effect to this legitimate expectation will infringe the accrued rights of any member of the public, or that this legitimate expectation is somehow outweighed by an overriding national or public interest.

148 To recapitulate, the parties largely agree that the political package – specifically, the representations made by AG Wong in 2018 – has a bearing on the constitutionality of s 377A, and that these representations are legally significant. We have identified the doctrine of substantive legitimate expectations as the basis on which these representations can be said to have legal force, albeit to the limited extent to which we have recognised that doctrine. The final question that remains is whether the intent underlying AG Wong’s representations would be sufficiently realised if we confined ourselves to the legitimate expectation identified at [146] above, namely, that the PP will not prosecute conduct falling within the Subset under s 377A.

149 We preface our decision on this point by stressing that we are relying on the doctrine of substantive legitimate expectations only in so far as the Anterior Question (as framed at [116] above) is concerned. With this in mind, we answer the Anterior Question by holding that s 377A is unenforceable in its entirety, unless and until the AG of the day provides clear notice that he, in his capacity as the PP: (a) intends to reassert his right to enforce s 377A proactively by way of prosecution; and (b) will no longer abide by the representations made by AG Wong in 2018 as to the prosecutorial policy that applies to conduct falling within the Subset. In holding as we do, we do not overlook the difficulty of the issues that were raised in Starkstrom.

150 It naturally flows from our holding that prosecutions under provisions such as ss 119 and 176 of the PC should not be instituted where the underlying offence is one under s 377A. In the same vein, offences under s 424 of the CPC should not be prosecuted where the “arrestable offence” (as statutorily defined) is one under s 377A. However, nothing in our holding affects the right of the police to investigate all conduct, including any conduct falling within the Subset and/or amounting to an offence under s 377A (see [113] above). Nor does our holding constrain the PP’s right to prosecute conduct falling outside the Subset where such conduct violates any other law, or impact the duties applicable to others arising, for instance, from their awareness of or participation in such conduct, whether actual or intended.

151 One might question why s 377A should be held to be unenforceable in its entirety even though the actual expectation that has arisen is only to the effect that no prosecutions will be brought in respect of conduct falling within the Subset (see [146] above). To this end, we first highlight that our finding is of modest consequence since it remains the prerogative of the AG of the day to decide and declare that he will no longer follow the prosecutorial policy articulated by AG Wong in 2018. Further, to the extent that the protection we accord somewhat exceeds the precise ambit of AG Wong’s representations, there is nothing to indicate that the additional aspects covered by our order runs contrary to AG Wong’s actual intentions as the PP. Rather, it seems that they were simply not within his focus at the time (see [107] above). If we are mistaken on this, as we have just noted, it remains open to AG Wong to indicate with reasonable notice that he does after all intend to exercise the AG’s prosecutorial prerogative in relation to the additional matters covered by our holding. Until such time, our holding would have the effect of:

(a) providing homosexual men with the full measure of accommodation contemplated by the Government and expressed by the Prime Minister during the s 377A Debates;

(b) minimising the prevailing legal untidiness and avoiding most of the uncertainties that would persist if we were to adopt a more restricted solution, as we explain at [152] below; and

(c) preserving the legislative status quo on s 377A and reserving the matter of its retention or repeal for further consideration by the Government and Parliament at an appropriate time.

152 It follows that our holding is not legally insignificant. Instead, it gives legal effect to AG Wong’s representations without importing the uncertainties that would otherwise continue to plague homosexual men (see [97]–[100] and [103]–[108] above). Such uncertainties are fundamentally antithetical to the rule of law, as we earlier noted (see [109] above). It will also be recalled that a central tenet of the Government’s position was the desire to ensure that homosexual men could live peacefully, free from harassment, in our society. As we alluded to at [114(d)] above, the legal uncertainties that homosexual men would continue to face unless s 377A is held to be unenforceable in its entirety (subject to the qualifications set out at [149]–[150] above) might fairly be regarded as giving rise to a sense of harassment, and must therefore be minimised if the political package is to be given full legal effect. Our finding therefore gives practical legal effect to both the political compromise on s 377A that the Government struck in 2007 and the legitimate expectation engendered by AG Wong’s representations.

153 Given our finding that the legal effect of AG Wong’s representations is that s 377A may not be enforced in its entirety unless and until the AG of the day signals a change in the prosecutorial policy, the appellants cannot be said to face any real and credible threat of prosecution under this provision at this time. It follows that there is in fact no controversy and, at present, no threat of any violation of their rights under Arts 9, 12 and 14 (see Tan Eng Hong (Standing) at [112]). The appellants hence do not have standing to mount the present constitutional challenges against s 377A. Although we nevertheless set out below our views on the constitutional issues that they have raised, we emphasise that our analysis of those issues is purely obiter since they do not arise for our determination in these appeals.

154 We end this section by reiterating that our recourse to the doctrine of substantive legitimate expectations should be regarded as wholly exceptional –necessitated only by the imperative of according legal effect to AG Wong’s representations, and a consequence of the congruent positions of Parliament, the Government and the AG. Section 377A has not been repealed, but neither can it be enforced. In these unique circumstances, the thorny issues surrounding the acceptance of the doctrine of substantive legitimate expectations as part of Singapore law do not arise, but they will have to be considered in a suitable case in the future.

155 We now consider Mr Thuraisingam’s and Mr Ravi’s argument on the immutability of sexual orientation. Relying on the expert evidence adduced in the proceedings below, they argue that sexual orientation cannot be wilfully changed and is not caused or influenced by socio-environmental factors. On this basis, they submit that s 377A effectively criminalises a class of persons for their sexual identity and is thus absurd and not “law” within the meaning of Art 9(1).

156 Despite Mr Thuraisingam’s best efforts to persuade us otherwise, we are not satisfied that it is properly within the court’s remit to decide whether sexual orientation is immutable. It bears emphasising that Dr Tan and Mr Ong, the appellants in CA 54 and CA 55 respectively, are not urging the court to find that their sexual orientation is immutable. Rather, they are urging us to hold that sexual orientation is immutable as a matter of general scientific fact. This is not a holding that we are entitled to make, for the court is concerned not with the laws of nature but with man-made laws to govern society.

157 As was stated in Mohammad Faizal bin Sabtu at [27], the judicial function:

The court’s role, put simply, is to pronounce the rights of the parties inter se. However, the scientific question which Dr Tan and Mr Ong have brought before us is not confined to their legal rights or the dispute at hand, and instead has wide-ranging implications across society. Moreover, the question whether sexual orientation is immutable calls for the application of the scientific method; it is not a question amenable to judicial resolution, having regard to the limited methods, tools or standards that are properly at our disposal (see Nagaenthran a/l K Dharmalingam v Public Prosecutor and another appeal [2019] 2 SLR 216 at [60] and [65]).

158 Before us, Mr Thuraisingam indicated that he was seeking “recognition of just one small point: [c]an sexual orientation be voluntarily changed or not?” Although he submitted that this was “a very narrow question”, we respectfully disagree. The breadth of the question framed by Mr Thuraisingam raises an alarm and quite clearly indicates that the question is in fact an extra-legal scientific matter beyond the court’s purview. We reiterate that the court is only equipped to make findings of fact that pertain specifically to the case brought before it, and not to make sweeping pronouncements of scientific fact.

159 In any event, it is unnecessary for us to decide whether sexual orientation is immutable because this question is not determinative of whether s 377A is constitutional. As Mr Thuraisingam acknowledged before us, even if we were to find that sexual orientation cannot be changed as a matter of volition, it would not necessarily follow that s 377A must be unconstitutional. Any argument to the effect that the Government can never regulate against immutable characteristics would run into obvious difficulties; indeed, Mr Thuraisingam disavowed any suggestion that he was mounting such an argument. To briefly illustrate why any such argument would be plainly untenable, the scientific literature on paedophilia suggests, to some extent, that paedophilia may be genetic and that paedophilic attractions may be immutable. Yet, no one would deem laws which criminalise sex with minors, such as ss 375(1)(b) and 376A(1) of the Revised PC, absurd or unconstitutional on the basis that they criminalise paedophiles for their identity. Similarly, kleptomania is a condition that, to some degree, may not be susceptible to volitional change, but there is simply no basis for asserting that laws which criminalise theft (such as ss 379 and 380 of the Revised PC) are absurd and therefore not “law” within the meaning of Art 9(1). Whether sexual orientation is immutable thus has little bearing in the final analysis on whether s 377A falls afoul of Art 9(1).

160 We therefore decline to make a finding on whether sexual orientation is immutable. Consequently, there is no need for us to delve into the expert evidence adduced by the parties. We only add that we were not particularly impressed by the quality of the expert evidence adduced on both sides. Among other things, the experts appeared to conflate sexual attraction with sexual behaviour at various points and at times used the term “choice” rather imprecisely when discussing sexual orientation.

161 We turn now to the proper interpretation of s 377A. The starting point in statutory interpretation is s 9A(1) of the Interpretation Act (Cap 1, 2002 Rev Ed) (“the IA”), which mandates a purposive approach to statutory interpretation. Section 9A(2)(a) of the IA provides that the court may consider extraneous material to confirm that the meaning of a provision is the ordinary meaning conveyed by the text of the provision. Pursuant to s 9A(2)(b) of the IA, extraneous material may be used to ascertain the meaning of a provision only if the provision is ambiguous or obscure, or if the ordinary meaning conveyed by the text of the provision, taking into account the statutory context and the legislative purpose of the written law, leads to a manifestly absurd or unreasonable result.

162 The parties agree that s 377A should be interpreted with reference to the Tan Cheng Bock framework, which is as follows (see Tan Cheng Bock at [37]):

(a) First, the court should ascertain the possible interpretations of the provision in question, having regard not just to the text of the provision but also to the context of that provision within the written law as a whole.

(b) Second, the court should ascertain the legislative purpose or object of the statute.

(c) Third, the court should compare the possible interpretations of the text against the purposes or objects of the statute.

163 Mr Singh’s submissions on the proper interpretation of s 377A are twofold. First, he argues that s 377A was intended to cover only non-penetrative sex acts and not penetrative sex acts as well, as the latter were already criminalised under s 377 when s 377A was enacted in 1938. He submits that while a “superficial linguistic interpretation of the words ‘grossly indecent’” [emphasis in original omitted] might lead one to conclude that penetrative sex acts fall within the scope of s 377A, a purposive interpretation of that provision would yield the opposite conclusion. Mr Singh further contends that no purpose would have been served in having s 377A criminalise penetrative sex acts when s 377 already punished the same conduct more harshly. Moreover, when s 377A was enacted in 1938, the Legislative Council could not have anticipated the eventual repeal of s 377 such that there would have been a need for s 377A to overlap with s 377 as a fallback. He thus urges us to find that s 377A was intended to criminalise only acts of gross indecency between men that did not amount to unnatural offences under s 377, such that s 377A would not extend to penetrative sex acts.

164 Mr Singh’s second argument, which Mr Ravi adopts, is that the legislative purpose of s 377A was to suppress male prostitution. In this regard, he relies on extraneous material to illustrate that the colonial authorities had been vexed by the problem of rampant male prostitution in the period leading up to the enactment of s 377A. Mr Singh submits that since the real mischief targeted by s 377A was male prostitution, the provision is over-inclusive in its criminalisation of non-commercial sex acts and therefore violates Art 12.

165 Before considering Mr Singh’s arguments and interpreting s 377A in accordance with the Tan Cheng Bock framework, we observe in passing that these arguments are internally inconsistent. According to Mr Singh, the extraneous material that he has put forth suggests a concern with curbing male prostitution. If we were to accept that interpretation of the extraneous material, the suppression of male prostitution, by way of the enactment of s 377A, must surely entail the criminalisation of penetrative sex acts. However, the contextual interpretation of s 377A that Mr Singh urges us to adopt suggests that this provision was not concerned with penetrative sex acts at all. Apart from the fact that these conclusions are contradictory, neither of them seems to us to be correct, as we explain below.

166 The first stage of the Tan Cheng Bock framework requires us to ascertain the possible interpretations of s 377A, having regard to its text and its context within the written law (that is to say, the PC) as a whole. To recapitulate, s 377A provides as follows:

167 In our judgment, the only possible interpretation of the words “gross indecency” is one that: (a) includes penetrative sex acts; and (b) is not limited to male prostitution. On a plain reading, the term “gross indecency” cannot sensibly be limited to non-penetrative sex acts and must extend to penetrative sex acts. We echo our observation in Lim Meng Suang (CA) (at [133]) that “[a]ny other interpretation would be illogical since it cannot be denied that acts of penetrative sex constitute the most serious instances of the possible acts of ‘gross indecency’” [emphasis in original]. Nor is the ordinary meaning of the term “gross indecency” limited to male prostitution – nothing in the language of s 377A suggests such a meaning. Had Parliament intended to significantly circumscribe s 377A by excluding penetrative sex acts or non-commercial sex acts from the acts of gross indecency criminalised thereunder, one would have expected Parliament to have done so expressly.

168 We pause here to address Mr Singh’s submission that the “written law” that should be considered at the first stage of the Tan Cheng Bock framework includes the MOO. He argues that s 377A should be interpreted in the context of both s 23 of the MOO and s 377, and that such an interpretation unambiguously points to the conclusion that s 377A excludes penetrative sex acts. With respect, this submission flies in the face of our decision in Tan Cheng Bock, which made it clear (at [37(a)]) that only the statute in which the provision in question is contained (in this case, the PC) constitutes relevant context at the first stage. The MOO might become relevant, if at all, at the second stage – that is to say, when ascertaining the legislative purpose or object of the PC. We add that the references in ss 9A(1) and 9A(2) of the IA to “the interpretation of a provision of a written law” [emphasis added] also make it clear that the “written law” that is relevant at the first stage of the Tan Cheng Bock framework is confined to the statute in which the provision in question is contained.

169 Mr Singh’s reliance on s 377 to interpret s 377A at the first stage of the Tan Cheng Bock framework is, with respect, similarly misplaced. While s 377A should be interpreted with regard to not only its text but also its context within the PC, s 377 was repealed in 2007. In other words, the PC that forms the relevant context when interpreting s 377A today excludes s 377.

170 Having regard to the text of s 377A and its context within the PC as a whole, we conclude that the ordinary meaning of the term “gross indecency” unambiguously includes penetrative sex acts and is not limited to male prostitution. With this in mind, we turn to the second and third stages of the Tan Cheng Bock framework.

171 The second stage of the Tan Cheng Bock framework requires the court to ascertain the legislative purpose or object of the specific provision in question (namely, s 377A) and of the part of the statute (in this case, the PC) in which the provision is situated. The court then moves to the third stage and compares the possible interpretations of the provision against the purpose or object of the relevant part of the statute. An interpretation which furthers the purpose or object of the written text should be preferred to an interpretation which does not (see Tan Cheng Bock at [37(c)] and [54(c)]).

172 We emphasise that the legislative purpose or object of a statutory provision should ordinarily be gleaned from the text of the provision in its statutory context. Primacy ought to be accorded to the text of the provision and its statutory context over any extraneous material (see Tan Cheng Bock at [43] and [54(c)(ii)]).

173 Section 377A criminalises acts of “gross indecency” between “male person[s]”. The broad way in which s 377A is framed militates against Mr Singh’s arguments that it criminalises only non-penetrative sex acts and that it was targeted at male prostitution (see also [167] above). Furthermore, although s 377A is listed under the section headed “Sexual offences” in Chapter XVI of the PC (which covers “Offences affecting the human body”), it is situated between the offences of sexual penetration of a corpse and sexual penetration with a living animal. What the latter offences criminalise are not assaults on bodily integrity per se, but the affronts to societal morality that the prohibited acts constitute. This suggests that the legislative purpose of the specific part of the PC in which s 377A is situated is to uphold public morality. Based on the text and the context of s 377A, we agree with the Judge that the legislative purpose of s 377A was likewise to safeguard public morals (see the Judgment at [146(d)]). The extraneous material before us did not persuade us otherwise, for the reasons that we explain below.

174 Before analysing the extraneous material before us, we make the preliminary observation that this material is of limited relevance and utility for two reasons. First, since we have found that the plain meaning of the term “gross indecency” clearly includes penetrative sex acts and is not limited to acts involving male prostitution (see [167] and [170] above), the extraneous material can only be used to confirm, and not to alter or supplant, that plain meaning (see s 9A(2)(a) of the IA and Tan Cheng Bock at [47]–[51]). As we emphasised in Tan Cheng Bock at [50], purposive interpretation is not an excuse for rewriting a statute; judicial interpretation is generally confined to giving a statutory provision a meaning that its language can bear. It is for this reason that extraneous material may be used to ascertain the meaning of a provision where the provision is ambiguous or obscure (see s 9A(2)(b)(i) of the IA), but not to override the express language of the provision, save in the very limited circumstances identified in s 9A(2)(b)(ii) of the IA.

175 We highlighted this point to Mr Singh at the hearing. In response, he argued that it was ambiguous whether s 377A was intended to overlap with s 377 and, consequently, whether s 377A criminalises both penetrative and non-penetrative sex acts. On that basis, he contended that it was permissible to consider the extraneous material in order to ascertain the meaning of s 377A.

176 With respect, it is untenable for Mr Singh to argue, on the one hand, that s 377A was clearly intended to criminalise only non-penetrative sex acts between men (as he did in his written submissions), and to contend, on the other hand, that recourse to extraneous material is permissible because of the apparent ambiguity as to whether s 377A was intended to cover both penetrative and non-penetrative sex acts (as he did in his oral submissions). These two positions are mutually exclusive. It seemed to us that Mr Singh’s primary position was the former and that the latter position was only adopted to admit extraneous material to vary the plain meaning of s 377A. In our view, any ambiguity in the ordinary meaning of s 377A is more imagined than real. As we pointed out to Mr Singh, in order to find the ambiguity in the meaning of s 377A that he contended for, one would have to find that a possible interpretation of s 377A, based on the express language of the provision, is that the provision covers only non-penetrative sex acts. If that were so, the extraneous material could then arguably assist the court when it compares the various possible interpretations of the text of s 377A. That, however, is not the case. Indeed, the plain language of s 377A pulls in the very opposite direction, as we have shown.

177 Second, it appears to us that the court should be especially cautious when relying on extraneous material in cases such as the present that involve legislation pre-dating s 9A of the IA. Section 377A was enacted at a time when recourse to extraneous material, whether legislative or non-legislative, would have been wholly unthinkable when discerning the legislative purpose of a statutory provision. Put another way, if a person had been prosecuted for sodomy under s 377A in 1938 or even in the decades thereafter, he would not have been in a position to challenge that prosecution by adducing extraneous material to dispute the legislative purpose of s 377A. Not only was at least some of such material unavailable until its declassification many years later, it was surely also beyond the contemplation of the Legislative Council in 1938 that such material, which pre-dated the enactment of s 9A of the IA and the introduction of the Tan Cheng Bock framework, could even be relied upon to ascertain the legislative purpose of s 377A. The fine-grained analysis of extraneous material is therefore entirely artificial in the context of legislation that pre-dates the legislative amendments and legal framework that permit recourse to extraneous material in the first place. It follows that non-legislative extraneous material is of even more limited utility (if any at all) to the task at hand, which is to ascertain the legislative purpose of s 377A.

178 Quite apart from the caution that we have just raised, there is another reason why non-legislative material, in particular, would generally be unhelpful in shedding light on the parliamentary intention behind a statutory provision. We echo the observations made in Tan Cheng Bock (at [52] and [54(c)(iv)]) that in deciding whether to consider extraneous material at all (and, if so, what weight to place on it), the relevant considerations are: (a) whether the material is clear and unequivocal; (b) whether it discloses the mischief targeted by or the legislative intention underlying the statutory provision in question; and (c) whether it is directed at the very point of statutory interpretation in dispute. These considerations naturally mean that the court is primarily concerned with legislative material when it undertakes the task of ascertaining Parliament’s intention in enacting a statutory provision, a point that the Judge highlighted (see the Judgment at [52]).

179 Despite the limited relevance of the extraneous material before us, we nonetheless proceed to analyse that material, which Mr Singh relies on to buttress his arguments that s 377A criminalises only non-penetrative sex acts and that its legislative purpose was to combat male prostitution. We reiterate the inconsistency in these positions, which we have already explained at [165] above. The extraneous material comprises:

(a) the speech made by Attorney-General C G Howell (“AG Howell”) in the Legislative Council on 13 June 1938, in moving the Penal Code (Amendment) Bill that was published on 29 April 1938 (“the 1938 Penal Code (Amendment) Bill”) to its second reading (“AG Howell’s speech”);

(b) the section in the 1938 Penal Code (Amendment) Bill titled “Objects and Reasons” (“the Objects and Reasons”);

(c) the Annual Reports on the Organisation and Administration of the Straits Settlements Police and on the State of Crime (“Annual Crime Reports”) for the years 1934 to 1938;

(d) an addendum of unknown provenance (“the Addendum”) to a 1940 report titled “Prosecutions, The Malayan ‘Sexual Perversion’ cases” from the local authorities to Sir G Gater, Permanent Under-Secretary of State for the Colonies (“the Malayan Prosecutions Memo”);

(e) a report dated 24 March 1938 from Sir Shenton Thomas, the Governor and High Commissioner of the Straits Settlements, to the Secretary of State for the Colonies, concerning the resignation of one Mr H Moses from service (“the Moses Report”); and

(f) the minutes of the Executive Council meeting held on 18 May 1938 (“the ECM minutes”).

180 We first examine the legislative extraneous material, namely, AG Howell’s speech and the Objects and Reasons. It should be noted at the outset that these items were considered at length in Lim Meng Suang (CA), in which we concluded that s 377A was not limited to non-penetrative sex acts nor to male prostitution. Having reconsidered this material, we remain of the same view.

181 The relevant portion of AG Howell’s speech reads as follows:

182 The “clause 4” referred to in the above extract was cl 4 of the 1938 Penal Code (Amendment) Bill, which introduced s 377A into the 1936 PC. AG Howell’s speech also refers to the inadequate punishment provided for under the MOO. It is undisputed that this was a reference to the punishment prescribed by s 23 of the MOO, which read as follows:

183 Under s 23, a first-time offender was liable to a fine not exceeding 20 dollars or an imprisonment term of up to 14 days, while a repeat offender was liable to a fine not exceeding 25 dollars or an imprisonment term of up to three months. Section 23 criminalised (among other things) indecent behaviour and persistently soliciting or importuning for immoral purposes, but only if such conduct took place in public. We note that, on a plain reading of s 23, indecent behaviour and persistently soliciting or importuning for immoral purposes were distinct and separate offences. This was what was held in Lim Meng Suang (CA) (at [132]) and we see no reason to depart from that holding.

184 In his speech, AG Howell stated that “[a]s the law now stands, such acts can only be dealt with, if at all, under the [MOO], and then only if committed in public”. What is apparent is that s 377A was intended to target acts that were, at least in some circumstances, already offences under s 23. The enactment of s 377A therefore appeared to have been aimed at extending the acts that were prohibited under s 23 to acts in private, and at enhancing the then applicable penalties. The question that remains is which acts (that were already criminalised under s 23) s 377A was meant to target: (a) indecent behaviour; (b) persistently soliciting or importuning for immoral purposes; or (c) both.

185 To answer this, we need to consider AG Howell’s intimation of the Legislative Council’s desire to “bring [the law] into line with English Criminal Law”. It is undisputed that the “English Criminal Law” referred to by AG Howell was s 11 (UK) of the 1885 UK Act, which criminalised acts of gross indecency between men and from which cl 4 of the 1938 Penal Code (Amendment) Bill was derived. Section 11 (UK) read as follows:

Section 11 (UK) criminalised acts of gross indecency between men, whether committed in public or in private. In contrast, s 23 of the MOO criminalised indecent behaviour (among other things), but only if such behaviour occurred in public. Section 377A was therefore intended to “strengthen the law [in the Straits Settlements] and to bring it into line with English Criminal Law” by criminalising acts of gross indecency between men, whether committed in public or in private. It is for this reason that s 377A largely replicates the language of s 11 (UK).

186 We agree with the Judge that, given the articulated legislative intent to bring the law in the Straits Settlements “into line with English Criminal Law”, the Legislative Council intended for all acts of gross indecency that were covered by s 11 (UK) to fall within the scope of s 377A, such that a consistent set of laws would govern all such conduct in the United Kingdom and its then colony, Singapore (see the Judgment at [121]). This, then, begs the question whether s 11 (UK) criminalised both penetrative and non-penetrative sex acts or only non-penetrative sex acts – a question that does not lend itself to any easy answers, given that the purpose of s 11 (UK) remains obscure.

187 As was noted in Lim Meng Suang (CA) at [117], there was no parliamentary debate on the substance of s 11 (UK) prior to its enactment. Section 11 (UK) was introduced at the eleventh hour in the House of Commons during a late-night debate at the report stage of the Criminal Law Amendment Bill 1885, with few MPs in attendance (see United Kingdom, House of Commons, Parliamentary Debates (6 August 1885) vol 300 at cols 1397–1398 (Mr Henry Labouchere, MP for Northampton), http://hansard.millbanksystems.com/commons/1885/aug/06/consideration> (accessed 21 February 2022). It had nothing to do with the general purpose of the 1885 UK Act, which was stated in the preamble to the Act to be “the Protection of Women and Girls, the suppression of brothels, and other purposes”. The legislative purpose of s 11 (UK) is therefore none too clear.

188 In the absence of direct or clear evidence as to the legislative purpose of s 11 (UK), the Judge relied on three pieces of extraneous material to discern that purpose. These were: (a) the case of The King v Barron [1914] 2 KB 570 (“Barron”); (b) the transcripts of and other reference material relating to Oscar Wilde’s trials; and (c) the United Kingdom Report of the Committee on Homosexual Offences and Prostitution (Cmnd 247, 1957) (Chairman: Sir John Wolfenden) (“the Wolfenden Report”) (see the Judgment at [122]–[128]).

189 We deal first with Barron, a case in which the appellant was indicted on a charge of gross indecency under s 11 (UK) in respect of an offence involving sodomy. The Judge inferred from the use of s 11 (UK) to prosecute offences involving sodomy, in cases such as Barron, that the scope of the provision went beyond non-penetrative sex acts and male prostitution (see the Judgment at [123]–[125]). With respect, we disagree with the Judge in so far as we are not inclined to rely on Barron to ascertain the legislative purpose of s 11 (UK) and, in turn, that of s 377A. We take this position for two reasons.

190 First, as we mentioned earlier (see [172] above), the legislative purpose of a statutory provision should ordinarily be discerned from its text and its context within the written law in which it is contained. While regard may be had to relevant extraneous material, case law as decided by the Judiciary should be scrutinised with utmost care to ascertain whether it in fact sheds any light on the Legislature’s intention in enacting the statutory provision in question. The possibility that the courts in the United Kingdom interpreted s 11 (UK) in a manner that was in fact inconsistent with the legislative intention underlying its enactment cannot be precluded. Moreover, because the way in which the prosecuting authorities deployed s 11 (UK) may not necessarily have corresponded to the legislative intention behind its enactment, we are unwilling to place weight on cases such as Barron in deciphering the legislative purpose of s 11 (UK).

191 Second, even if we were to accept that the courts’ interpretation of a statutory provision may, in some cases, be of use in illuminating the legislative purpose of that provision, it bears emphasising that Barron concerned a different statute and one of unclear pedigree – namely, s 11 (UK), and not s 377A. While AG Howell’s speech suggests an endeavour to bring the law in the Straits Settlements into line with that in England, the reasons for such an endeavour by the Legislative Council may be quite different from why s 11 (UK) was introduced in England. The same point was underscored in Lim Meng Suang (CA) at [118], where we highlighted that s 377A had been enacted some 53 years after the inception of s 11 (UK) in England. The treatment of s 11 (UK) in Barron is therefore of no assistance to our present inquiry into the legislative purpose of s 377A.

192 We add that, in any event, Barron appears to contradict Mr Singh’s position that s 11 (UK) and, hence, s 377A criminalise only non-penetrative sex acts. The appellant in Barron had previously been indicted for sodomy and had been acquitted. He was then indicted for committing an act of gross indecency on the same facts as those upon which the indictment for sodomy had been founded. His plea of autrefois acquit was overruled, whereupon he pleaded guilty to the gross indecency charge. Thereafter, he appealed on the basis that the judge had erred in rejecting his plea of autrefois acquit.

193 The English Court of Criminal Appeal dismissed the appeal. The court started from the premise (at 574) that the principle of autrefois acquit did not permit a person to be twice in peril of being convicted of the same offence. It then noted (at 576) that the appellant could not have been convicted of gross indecency at the first trial. It also observed that “sodomy involves gross indecency and something else”, and that penetration was an essential element of sodomy but not of gross indecency. Finally, it concluded that the offence to which the appellant had pleaded guilty at the second trial (namely, gross indecency) was not the same as that of which he had been acquitted at the first trial (namely, sodomy).

194 It is unclear to us why the appellant could not have been convicted of gross indecency at the first trial despite having been acquitted of sodomy. Mr Singh argues that this was because the offences of sodomy and gross indecency did not overlap; therefore, the jury could not have convicted the appellant of gross indecency at the first trial on the facts upon which his indictment for sodomy had been founded. This, however, is incorrect. As we noted above, the English Court of Criminal Appeal held that “sodomy involves gross indecency and something else” [emphasis added]. In our view, this suggests that the two offences overlap and that sodomy is an aggravated form of gross indecency. On the other hand, Ms Tan submits that the reason why the jury could not have convicted the appellant of gross indecency at the first trial is a jurisdictional rule at common law that two indictments cannot be the subject of the same trial. The arguments proffered by the parties do not provide particularly strong support either way. Instead, what is crucial is that the appellant was convicted of gross indecency at the second trial, on the very same facts which underpinned his indictment for sodomy at the first trial. This shows beyond peradventure that penetrative sex acts could constitute acts of gross indecency under s 11 (UK), and we see no reason why the position under s 377A should be any different.

195 As regards the transcripts of and other reference material pertaining to Oscar Wilde’s trials, we first note that none of the counsel for the appellants rely on this material. The Judge observed that Wilde faced multiple charges of gross indecency under s 11 (UK), and that evidence of his having committed sodomy was adduced in relation to some of those charges. While the first trial against Wilde resulted in a hung jury, he was convicted at the second trial of seven counts of gross indecency. For much the same reasons as those stated at [190]–[191] above, we decline to place any weight on the material relating to Wilde’s trials. Indeed, Mr Singh himself submits that this material is irrelevant in determining the scope of s 11 (UK).

196 In any event, while the juries in both of Wilde’s trials did not provide any reasons for their decisions, what is clear is that at least two of Wilde’s seven convictions for gross indecency were based on his having committed sodomy, rather than any non-penetrative sex acts, with one Charles Parker. The material relating to Wilde’s trials, much like Barron, thus confirms that penetrative sex acts could amount to acts of gross indecency under s 11 (UK). This undermines Mr Singh’s argument that s 377A criminalises only non-penetrative sex acts.

197 We turn next to the Wolfenden Report. This report was prepared by the Departmental Committee on Homosexual Offences and Prostitution (“the Wolfenden Committee”), which was formed by the United Kingdom Government to comprehensively review homosexual offences in the United Kingdom and to recommend legislative changes. In particular, the Wolfenden Committee considered arguments for and against the abolition of the offence of gross indecency. It observed (at para 104 of the Wolfenden Report) that gross indecency was “not defined by statute” but “appear[ed] … to cover any act involving sexual indecency between two male persons”, an observation that the Judge adopted (see the Judgment at [122]).

198 Ms Tan highlights the Wolfenden Committee’s observation (at para 105 of the Wolfenden Report) that the offence of gross indecency usually took one of three forms, one of which was “oral-genital contact”, meaning penetrative oral sex. In other words, s 11 (UK) appeared to cover penetrative sex acts. Ms Tan adds that the Wolfenden Report does not cite the suppression of male prostitution as a reason for the retention of the offence of gross indecency, which puts paid to any suggestion that the purpose of s 377A was limited to the curbing of male prostitution.

199 According to Mr Singh, however, the Wolfenden Report shows that the offence of gross indecency (under s 11 (UK)) did not overlap with the offence of sodomy; accordingly, s 377A does not overlap with s 377. His argument can be summarised as follows:

(a) Paragraph 107 of the Wolfenden Report states that s 11 (UK) “merely extended to homosexual indecencies other than buggery the law which previously applied to buggery” [emphasis added in italics and bold italics]. Mr Singh contends that this paragraph shows that s 11 (UK) was not intended to cover sodomy.

(b) Buggery, which was originally criminalised under the Buggery Act 1533 (c 6) (UK) and later criminalised as sodomy under s 61 of the Offences Against the Person Act 1861 (c 100) (UK) (“the UK OAPA”), meant anal penetration only; it did not include oral penetration. The fact that the Wolfenden Report states that gross indecency under s 11 (UK) covered oral-genital contact, but omits to state that it covered sodomy, shows that sodomy and gross indecency were non-overlapping offences.

(c) The fact that s 11 (UK) covered oral penetration does not buttress Ms Tan’s case. Although sodomy (under s 61 of the UK OAPA) meant anal penetration only, s 377 had been interpreted to cover both oral penetration and sodomy by the time s 377A was enacted. As mentioned, sodomy (under s 61 of the UK OAPA) and gross indecency (under s 11 (UK)) were non-overlapping offences since only the latter covered oral penetration. In the same vein, the offences under ss 377 and 377A did not overlap – only the former covered oral penetration.

200 Curiously, neither Mr Singh nor Ms Tan referred to the following paragraphs of the Wolfenden Report, which we find enlightening:

The aforesaid paragraphs of the Wolfenden Report speak with one voice – buggery and attempted buggery were acts of gross indecency, which explains the reference to s 11 (UK) criminalising “other” acts of gross indecency. In the same vein, para 88 of the Wolfenden Report discusses whether there was even a need to retain buggery as a separate offence when it was but one of many forms of indecent assault or gross indecency. The Wolfenden Report thus confirms that the acts of gross indecency criminalised under s 11 (UK) include penetrative sex acts such as buggery (or, in other words, sodomy).

201 It appears that Mr Singh misunderstood the sentence in para 107 of the Wolfenden Report – namely, that s 11 (UK) “merely extended to homosexual indecencies other than buggery the law which previously applied to buggery” [emphasis added in italics and bold italics] – to mean that s 11 (UK) covered acts of gross indecency except for buggery (see [199(a)] above). On the contrary, it is clear from our foregoing analysis that the Wolfenden Committee intended to convey that s 11 (UK) was a widening provision. Section 11 (UK) expanded the law which had previously applied to buggery by creating a wider offence that criminalised both buggery and acts of gross indecency falling short of buggery (which had not hitherto been criminalised). Furthermore, and unlike what Mr Singh suggests, it is immaterial that s 11 (UK) did not explicitly cover sodomy. There was simply no such need because sodomy could simultaneously be a stand-alone offence under s 61 of the UK OAPA and fall under the less serious but wider offence of gross indecency under s 11 (UK).

202 We note that the Wolfenden Report states (at para 105) that the offence of gross indecency under s 11 (UK) “usually [took] one of three forms: either … mutual masturbation; or … some form of intercrural contact; or oral-genital contact”. Although the last of these is a form of penetrative sex, nothing much turns on this. Simply put, the fact that the offence of gross indecency under s 11 (UK) “usually” assumed the form of these sex acts does not mean that that was invariably or necessarily the case, or that the offence did not or could not extend to other forms of penetrative and non-penetrative sex acts. Hence, even though Ms Tan and Mr Singh each submitted on the significance of the fact that s 11 (UK) appeared to cover oral penetration, we place little weight on those submissions.

203 In summary, while we are unable to draw a firm conclusion as to the legislative purpose of s 11 (UK), there is on balance more support for the position that the provision covered penetrative sex acts. Before us, the parties did not expend much energy or time on the legislative purpose and proper interpretation of s 11 (UK). We thus say no more on s 11 (UK), beyond noting that nothing in the foregoing analysis alters our finding that s 377A covers both penetrative and non-penetrative sex acts.

204 We next examine the Objects and Reasons, which, according to Mr Singh, show that s 377A criminalises only non-penetrative sex acts. The relevant part of the Objects and Reasons reads as follows:

Section 377 criminalised penetrative sex acts, whether committed in public or in private, and whether consensual or non-consensual. Mr Singh argues that it is “crystal clear” from the Objects and Reasons that s 377A does not cover penetrative sex acts, which fell within the scope of s 377 instead.

205 We respectfully disagree. In our view, the Objects and Reasons do not lead inexorably to the conclusion that Mr Singh contends for. The Objects and Reasons state that s 377A “makes punishable acts of gross indecency between male persons which do not amount to an unnatural offence within the meaning of section 377”. Acts of gross indecency could, however, be made punishable in two ways: (a) by creating a new offence which criminalised only non-penetrative sex acts (such that there would be no overlap between ss 377 and 377A); or (b) by creating a new offence that covered both penetrative and non-penetrative sex acts (such that penetrative sex acts would be covered by both ss 377 and 377A). The plain language of both the Objects and Reasons and s 377A itself is wide enough to support conclusion (b), and there is nothing to suggest that what the Legislative Council intended was conclusion (a).

206 Further, if s 377A was intended to cover only non-penetrative sex acts, as Mr Singh submits, AG Howell’s allusion to the need to supplement s 23 of the MOO would be puzzling. Mr Singh’s argument would mean that the Legislative Council had intended to supplement s 23 – which criminalised both penetrative and non-penetrative sex acts in public (as indecent behaviour) – by criminalising only non-penetrative sex acts committed in public or in private, to the exclusion of penetrative sex acts committed in private. Even though penetrative sex acts committed in private were already offences under s 377 at the time that s 377A was enacted, AG Howell spoke of the need to supplement s 23 without referring to s 377 at all. The fact that penetrative sex acts committed in private were already criminalised under s 377 therefore does not entail the conclusion that s 377A was intended to criminalise only non-penetrative sex acts committed in private: AG Howell simply highlighted the need to supplement s 23, without suggesting that any overlap with s 377 was to be avoided.

207 We also note that while AG Howell’s speech referred to the need to supplement the MOO, the Objects and Reasons speak of the need to supplement s 377. However, any inconsistency between AG Howell’s speech and the Objects and Reasons, in so far as they referred to the need to supplement different statutes, is more apparent than real. Turning first to AG Howell’s speech, it will be recalled that s 23 criminalised “indecent behaviour” (which would include, among other things, penetrative sex acts), but only if such behaviour occurred in public. Section 377A was thus intended to supplement s 23 by criminalising acts of gross indecency committed in private. The Objects and Reasons, on the other hand, articulate a need to supplement s 377, which criminalised penetrative sex, whether committed in public or in private. Section 377A therefore simultaneously supplemented s 377 by criminalising acts of gross indecency which fell short of penetrative sex. Read harmoniously, AG Howell’s speech and the Objects and Reasons support the construction that s 377A was intended to criminalise both penetrative and non-penetrative sex acts, whether committed in public or in private. This was also the conclusion which was reached in Lim Meng Suang (CA), and which we reaffirm:

208 Mr Ravi claims that the acts that “[could] only be dealt with, if at all, under the [MOO], and then only if committed in public”, which AG Howell’s speech referred to, concerned only prostitution. According to Mr Ravi, since s 23 of the MOO criminalised acts of prostitution in public, “[t]he clear intention was for section 377A to strengthen the Police’s ability to deal with all forms of prostitution by criminalising private instances of male prostitution as well”. He therefore asserts that s 377A was enacted to suppress male prostitution. However, this assertion blatantly ignores the fact that s 23 criminalised not only persistently soliciting or importuning for immoral purposes, but also indecent behaviour, which was a separate and distinct offence (see [183] above). To establish which limb of s 23 it was that s 377A aimed to strengthen, one must consider AG Howell’s express reference (in the very same speech) to the Legislative Council’s desire to bring the law in the Straits Settlements “into line with English Criminal Law”. As we have noted, the “English Criminal Law” that AG Howell spoke of was a reference to s 11 (UK) (see [185] above). While the origins and legislative purpose of s 11 (UK) are obscure, there is no evidence at all to suggest that it criminalised only homosexual sex acts with male prostitutes. Indeed, none of the appellants has submitted any material to support such a proposition.

209 Mr Ravi’s argument is also plainly inconsistent with AG Howell’s speech. It is clear from AG Howell’s speech that s 377A was aimed at (among other things) extending the acts prohibited under s 23 of the MOO from public to private acts, but in respect of the same type of acts. On Mr Ravi’s case, however, the acts purportedly targeted by s 377A were commercial sex acts between men, whereas the acts prohibited under s 23 were indecent behaviour and persistently soliciting or importuning for immoral purposes. Put another way, if Mr Ravi’s argument were accepted, the acts that “[could] … be dealt with, if at all, under the [MOO]” and the acts proscribed by s 377A would have been incongruent, contrary to AG Howell’s indications. Our interpretation, in contrast, coheres with AG Howell’s speech: the prohibited conduct under s 23 that AG Howell alluded to was indecent behaviour, while the acts prohibited under s 377A were acts of gross indecency. Both provisions targeted penetrative and non-penetrative sex acts between men, whether of a commercial or non-commercial nature. Where they differed was that the former prohibited only public acts, whereas the latter covered both public and private acts.

210 We add that neither AG Howell’s speech nor the Objects and Reasons even allude to male prostitution, which fortifies our conclusion that s 377A is a provision of general application and that acts of “gross indecency” encompass more than homosexual sex acts with male prostitutes.

211 Having considered AG Howell’s speech and the Objects and Reasons, neither of which provide a compelling case for Mr Singh’s and Mr Ravi’s position, we move on to the non-legislative extraneous material, beginning with the 1934 to 1938 Annual Crime Reports. We first highlight that the 1936 to 1938 Annual Crime Reports were previously considered by this court in Lim Meng Suang (CA) (at [125]–[127] and [142]). The 1934 and 1935 Annual Crime Reports, however, were not before us in Lim Meng Suang (CA). Mr Singh submits that the 1934 to 1938 Annual Crime Reports, when viewed in totality, illustrate that male prostitution was a problem which only emerged in the period leading up to the passage of the 1938 Penal Code (Amendment) Bill, and which that Bill was intended to address by introducing s 377A into the 1936 PC.

212 As a preliminary matter, we are of the view that little weight should be placed on the Annual Crime Reports. As we emphasised in Tan Cheng Bock (at [52]), not all extraneous material that can potentially touch on the purpose of the legislative provision in question is relevant. The Annual Crime Reports, which were prepared by the Straits Settlements police, may well have set out the social and/or historical backdrop to the introduction of s 377A, but they were not targeted at the specific issue of the legislative intention underlying s 377A (see Tan Cheng Bock at [52(c)] and [54(c)(iv)]).

213 Be that as it may, we nonetheless examine the 1934 to 1938 Annual Crime Reports individually for completeness. Quite to the contrary of the point which Mr Singh and Mr Ravi sought to demonstrate, the 1934 and 1935 Annual Crime Reports only mention female prostitution (under the heading “Social Services”) and omit any mention of male prostitution. Male prostitution is first mentioned in passing under the heading “Social Services” in the 1936 Annual Crime Report, as follows:

214 The 1937 Annual Crime Report likewise discusses male prostitution, albeit under the heading “Public Morals”:

215 Male prostitution similarly features in the 1938 Annual Crime Report, again under the heading “Public Morals”:

216 Mr Singh submits that the problem of male prostitution assumed noticeable urgency just before the enactment of s 377A, as can be seen from the 1937 Annual Crime Report. He further highlights that the passage of s 377A in the Legislative Council was underway when that report was published on 30 June 1938. He thus submits that s 377A must have been introduced to address the specific concern of male prostitution, and not to criminalise all sexual conduct between men. We first observe that Mr Singh’s submission runs into the difficulty that there is nothing in any of the references to male prostitution in any of the relevant Annual Crime Reports to suggest that the authorities felt hampered by a lack of legislative tools to deal with this issue. On the contrary, it appears that the issue was being dealt with effectively as and when it arose.

217 Aside from this, we consider with respect, that the 1934 to 1938 Annual Crime Reports do not show that s 377A was intended to have the limited purpose of curbing male prostitution. We make two observations in this regard. First, it is evident from the 1937 and 1938 Annual Crime Reports that male prostitution was discussed in the context of (and, indeed, under the heading) “Public Morals”. Furthermore, these two reports elaborate on how female prostitution was an affront to public morals, and raise male prostitution only fleetingly as part of a broad discussion on how prostitution could undermine public morals (see [214]–[215] above). In other words, the 1937 and 1938 Annual Crime Reports intimate an anxiety over the threat that prostitution, whether male or female, posed to public morals. They are hence consistent with s 377A having the wider legislative purpose of safeguarding public morality, which was the legislative purpose found by the Judge (see the Judgment at [146(d)]) and affirmed by us (see [173] above), rather than the narrower legislative purpose of combatting male prostitution specifically.

218 Second, our finding that s 377A is of general application is entirely consistent with the 1934 to 1938 Annual Crime Reports. Since the plain meaning of the phrase “act[s] of gross indecency” in s 377A encompasses sex acts (whether penetrative or non-penetrative) between men in general (see [167] and [170] above), acts relating to male prostitution would also be captured under s 377A, thereby alleviating the concerns over male prostitution that were raised in the 1936 to 1938 Annual Crime Reports. This was also the conclusion reached in Lim Meng Suang (CA) (at [142]). In our view, the 1934 and 1935 Annual Crime Reports, which are before us for the first time, do not affect that conclusion.

219 We now deal with the remaining extraneous material, which comprises: (a) the Addendum to the Malayan Prosecutions Memo; (b) the Moses Report; and (c) the ECM minutes.

220 Mr Singh relies on the Addendum to the Malayan Prosecutions Memo to bolster his case that s 377A was introduced to address the specific problem of male prostitution. The Malayan Prosecutions Memo was declassified by the United Kingdom Government in 2016. In summary, the Addendum depicts the unsatisfactory state of affairs regarding the prosecution, dismissal and/or removal from the Malayan Civil Service of civil servants who faced disciplinary action. It begins with an observation of an “outbreak” of “cases of this nature” in Malaya “[a]t the beginning of 1938” before detailing the cases of two colonial officials, Mr Reeves and Mr Rivaz. Mr Reeves had been strongly suspected of associating with “catamites”, which Mr Singh submitted was a reference to male prostitutes. Mr Reeves was charged under Colonial Regulation 68, but as none of the charges against him could be proved, he was not dismissed or removed from the Malayan Civil Service. Mr Rivaz too faced charges under Colonial Regulation 68. It was noted in the Addendum that “the charges against Mr. Rivaz have recently, by an amendment of the law, been made offences under the [1936 PC]”. Unlike Mr Reeves, Mr Rivaz was dismissed from the Malayan Customs Service as a number of charges against him were proved. Mr Singh submits that s 377A was clearly introduced to address the “outbreak” of male prostitution and the problem of civil servants patronising male prostitutes.

221 As for the Moses Report, which was sent by Sir Shenton Thomas, the Governor and High Commissioner of the Straits Settlements, to the Secretary of State for the Colonies, it details the circumstances that led to the resignation of one Mr H Moses, a “European Warder” at the Straits Settlements Prisons, from service. In essence, Mr Moses had been arrested in a hotel room while naked in bed with “two known catamites” and had been charged with attempted sodomy thereafter. According to one of the arresting officers, Mr Moses had admitted that if not for the police’s intervention, he would have committed the offence of sodomy. Mr Singh argues that the Moses Report, like the Addendum to the Malayan Prosecutions Memo, illustrates that the legislative purpose of s 377A was to combat male prostitution.

222 In our view, the Addendum to the Malayan Prosecutions Memo and the Moses Report are irrelevant to begin with and should not be considered. As non-legislative material, they do not and could not possibly disclose anything about the legislative intention underlying s 377A, and are thus not directed at the specific point of statutory interpretation in dispute (see Tan Cheng Bock at [52(c)] and [54(c)(iv)]).

223 In any event, both these pieces of extraneous material hardly suggest that the legislative object of s 377A was solely to combat male prostitution. The Addendum to the Malayan Prosecutions Memo provides no more than a snapshot of two instances of colonial civil servants associating with “catamites”. It cannot be surmised from these two instances that s 377A was introduced to deal only with cases involving “catamites”, even if we assume that that term indeed refers to male prostitutes. Further, even if the enactment of s 377A did criminalise the private conduct that Mr Rivaz had engaged in, it is simply untenable to say that such conduct was in breach of that provision only because he had engaged in it in the company of male prostitutes. Likewise, the Moses Report elaborates on a single case of a colonial civil servant engaging in homosexual sex acts with “catamites”. This can hardly constitute a basis for discerning the Legislative Council’s intention in enacting s 377A.

224 The ECM minutes are the last piece of extraneous material that Mr Singh seeks to rely on. They pertain to the Executive Council meeting held on 18 May 1938, shortly before the 1938 Penal Code (Amendment) Bill was passed, and detail the Executive Council’s deliberations over a memorandum by the police on “sexual perversion”. The salient portions of the minutes read as follows:

It is apparent from the date and the contents of the ECM minutes that “the Bill to amend the [1936 PC]” and “the clause relating to such offences” refer to the 1938 Penal Code (Amendment) Bill and cl 4 thereof respectively. According to Mr Singh, the ECM minutes reveal that s 377A was intended to deal with only non-penetrative sexual activity between men in private, that being the conduct which had not been criminalised as at 18 May 1938. He therefore argues that the Legislative Council did not intend the offences under ss 377 and 377A to overlap.

225 The ECM minutes pertain to the 1938 Penal Code (Amendment) Bill and what AG Howell was to say in his speech to the Legislative Council in moving the second reading of that Bill. As Mr Singh notes, nine out of the ten members present at the Executive Council meeting on 18 May 1938 were members of the Legislative Council. Most importantly, the ECM minutes shed light on the legislative intention underlying s 377A and touch on the very point of statutory interpretation in dispute (see Tan Cheng Bock at [52] and [54(c)(iv)]). We are therefore of the view that the ECM minutes merit greater consideration than the other non-legislative extraneous material that we have examined above.

226 Based on the ECM minutes, it may be argued, as Mr Singh does, that s 377A was aimed not at sodomy, which had already been criminalised prior to 1938, but at practices which were not offences then. As at 18 May 1938, no one could be prosecuted for the “practices” which s 377A targeted unless the act in question constituted an offence against public decency, in which case prosecution under s 23 of the MOO would be possible.

227 On this reading, the ECM minutes would seem to suggest that s 377A was intended to supplement the then existing law by criminalising non-penetrative sex acts between men in private, which were the only form of sexual conduct between men that had not already been criminalised as at 18 May 1938. Had s 377A been intended to cover both penetrative and non-penetrative sex acts, the statement in the ECM minutes that “the practices against which the new law is aimed are not, at present, offences against the law” [emphasis added] might not make sense since penetrative sex acts between men were already criminalised as sodomy under s 377. For the same reason, if penetrative sex acts fell within the scope of s 377A, the statement in the ECM minutes that “no person can be prosecuted for such practices until the amending Bill is passed, unless … the act constitutes an offence against public decency in which case prosecution can take place under the [MOO]” [emphasis added] would also be inexplicable. The ECM minutes further elaborate that “the clause relating to such offences” – that is to say, cl 4 of the 1938 Penal Code (Amendment) Bill (which introduced s 377A into the 1936 PC) – was intended to criminalise “certain practices” which “[had] not hitherto been made punishable in this country” [emphasis added]. Once again, only non-penetrative sex acts between men in private had not been made punishable as at 18 May 1938. This might also explain the clarification in the ECM minutes that “the amending law”, namely, the 1938 Penal Code (Amendment) Ordinance (see [16] above), would apply only prospectively and not retrospectively. On the other hand, as we note below, there is another perspective, which is that the enactment of s 377A might well have been aimed at non-penetrative sex acts between males in private, but even so, it does not necessarily follow that the section was intended to apply to only such acts. We will return to this shortly.

228 If s 377A was indeed intended to criminalise only non-penetrative sex acts between men in private, as the ECM minutes might appear on one reading to suggest, that would support Mr Singh’s argument that the offences under ss 377 and 377A were not intended to overlap. There are, however, two difficulties that stand in the way. We first reiterate the point we made at [177] above concerning the particular need for caution when resorting to extraneous material to discern the legislative intent underlying legislation that pre-dates the enactment of s 9A of the IA, which is the provision that permits recourse to such material in the first place. The danger is accentuated in the present context – to the extent that there was an express articulation of the legislative intent that led to the enactment of s 377A, it is to be found in AG Howell’s speech and the Objects and Reasons. For the reasons we have set out at considerable length at [181]–[207] above, these two pieces of extraneous material simply do not support Mr Singh’s contention.

229 The ECM minutes detail the deliberations of the Executive Council as to what AG Howell was to say at the second reading of the 1938 Penal Code (Amendment) Bill. The artificiality of the interpretive exercise that Mr Singh urges us to embark on is compounded when he contends that we should have regard to the ECM minutes because they ostensibly show that AG Howell’s speech was intended to convey a meaning that the text of that speech does not itself bear, in order then to construe the text of s 377A in a way that its language does not support. Mr Singh’s suggestion is even more untenable given that at the time s 377A was enacted, it would have been unthinkable for the court to have had regard to any such extraneous material. Whether a second reading speech is made in the expectation that it will inform the court’s understanding of the relevant legislation will inevitably have a bearing on the contents of that speech.

230 We return here to the point we alluded to earlier. In considering the ECM minutes, the same points that have been made in relation to AG Howell’s speech and the Objects and Reasons, as summarised above, have to be weighed in the balance. As we have already pointed out, the fact that cl 4 of the 1938 Penal Code (Amendment) Bill may have been “aimed” at non-penetrative sex acts between men in private, because that category of male homosexual sex acts had not hitherto been criminalised, does not mean that the clause could not or did not also cover other male homosexual sex acts. Whether the eventual enactment (that is to say, s 377A) did or did not extend beyond non-penetrative sex acts between men in private depends ultimately on its text, and nothing in the ECM minutes changes the analysis we have set out.

231 This leads to our second point: for the reasons explained at [161] and [174] above, the ECM minutes cannot vary the ordinary meaning of s 377A, which unambiguously includes both penetrative and non-penetrative sex acts.

232 One may ask whether there is any real value in considering the ECM minutes if they can only be used to confirm the ordinary meaning conveyed by the text of s 377A, even if they call that ordinary meaning into question (see s 9A(2)(a) of the IA). The limited use of extraneous material under s 9A(2)(a) of the IA was expressly considered in Tan Cheng Bock, where we explained:

233 In Public Prosecutor v Low Kok Heng [2007] 4 SLR(R) 183 (“Low Kok Heng”) at [50]–[52], which is cited in the above extract from Tan Cheng Bock, the High Court stressed (at [50]) the need for the court to assiduously guard against rewriting a statute in the name of adopting a purposive approach:

234 Hence, even if the ECM minutes might arguably lend some support to Mr Singh’s argument that s 377A was aimed exclusively at non-penetrative sex acts between men in private, that interpretation is ultimately at odds with the express language of s 377A and is therefore inadmissible. Our role is to interpret s 377A as enacted and not the ECM minutes, and there is no question that acts of “gross indecency” include penetrative sex acts. Moreover, and as we have already emphasised, it would be especially fanciful for us to accord any more weight to the ECM minutes when recourse to extraneous material would have been inconceivable at the time s 377A was enacted.

235 To conclude, the legislative purpose of s 377A was not to stamp out male prostitution but to safeguard public morals generally. This was also the conclusion that we arrived at in Lim Meng Suang (CA) (at [141]–[143]) after considering all the then available extraneous material. The new extraneous material before us does not displace that finding. The legislative purpose of s 377A – namely, the safeguarding of public morals generally – is also consonant with the legislative purpose of the part of the PC in which that provision is situated (see [173] above). Furthermore, the ordinary meaning of the phrase “gross indecency” (as held at [167] and [170] above) is perfectly consistent with and, in fact, furthers the legislative purpose of s 377A. It follows that there is simply no basis for preferring Mr Singh’s and Mr Ravi’s interpretation of “gross indecency” over the plain meaning of that phrase.

236 In that light, we now address Mr Ravi’s and Mr Thuraisingam’s argument that s 377A violates Art 9.

237 As a preliminary matter, we reiterate our finding at [149] above that s 377A is currently unenforceable in its entirety. The effect of this is that the appellants do not face any real and credible threat of prosecution under this provision at this time (see Tan Eng Hong (Standing) at [112], which we referred to at [153] above). There is hence no basis for Dr Tan and Mr Ong to invoke Art 9, and the question of whether s 377A is consistent with Art 9 is moot. Nonetheless, we consider the parties’ arguments on this issue for completeness. Our view, albeit in obiter, is that the Art 9 constitutional challenge is unmeritorious. Section 377A does not deprive a person of “life or personal liberty” in a sense that is caught by Art 9(1); nor can s 377A be said to be so absurd that it does not constitute a valid law for the purposes of Art 9(1).

238 Article 9(1) provides as follows:

239 To succeed in their Art 9 constitutional challenge, Dr Tan and Mr Ong must show that (see Yong Vui Kong v Public Prosecutor [2015] 2 SLR 1129 (“Yong Vui Kong (Caning)”) at [14]):

(a) s 377A deprives or threatens to deprive them of their right to “life or personal liberty” in a sense that is caught by Art 9(1); and

(b) s 377A is not a valid law for the purposes of Art 9(1) and is thus an unlawful deprivation of “life or personal liberty”.

240 Mr Thuraisingam and Mr Ravi advance overlapping arguments in support of their position that s 377A violates Art 9(1). These arguments largely target the validity of s 377A as a law that deprives a person of his personal liberty.

241 Mr Thuraisingam makes three principal arguments on this front. First, he contends that s 377A is “absurd” because it exposes a class of persons to the risk of incarceration on account of their sexual identity without a good and compelling reason. Second, he submits that s 377A is “arbitrary” because it fails the “reasonable classification” test. Third, he argues that s 377A is “contrary to the rule of law” because individuals who engage in conduct prohibited under that provision cannot reasonably foresee the legal consequences of their actions. He submits that although the 2018 AGC Press Release suggests that reported cases of conduct falling within the Subset (as defined at [146] above) will be investigated but will likely not be prosecuted, it remains unclear if and when such cases will result in prosecution. Mr Ravi aligns himself with Mr Thuraisingam and adds that the ambiguity surrounding the enforcement of s 377A generates uncertainty as to how other statutory provisions, such as ss 119 and 176 of the PC and s 424 of the CPC, ought to be construed.

242 The “arbitrariness” argument will be dealt with in the context of the “reasonable classification” test when we consider the Art 12 constitutional challenge. As for the rule of law concerns, these have been dealt with by way of our limited recognition of the doctrine of substantive legitimate expectations in the specific context of s 377A. We therefore focus on the claim that s 377A is absurd, which we analyse by reference to the following sub-issues:

(a) whether s 377A deprives a person of “life or personal liberty” in a sense that is caught by Art 9(1); and

(b) whether s 377A is “absurd” such that it is not a valid law for the purposes of Art 9(1).

243 Mr Thuraisingam and Mr Ravi implicitly urge us to recognise that sexual identity or, specifically, their clients’ homosexual identity, falls within the “life or personal liberty” that Art 9(1) protects. This is most evident from their following arguments:

(a) Mr Thuraisingam submits that “deprivation of life or personal liberty on the ground of identity can only ever be justified in the very rare case that doing so is necessary to advance a compelling state interest” [emphasis in original omitted; emphasis added in italics].

(b) Mr Ravi takes the position that s 377A criminalises a person’s sexual identity and argues that:

244 In our judgment, these arguments must be rejected because the deprivation of personal liberty that s 377A may engender does not fall within the scope of protection afforded by Art 9(1).

245 It is well established that unenumerated substantive rights cannot be read into the Constitution. The reasons for this were set out in detail in Yong Vui Kong (Caning) as follows:

These holdings have been repeatedly cited and affirmed in our local jurisprudence (see, for example, Chijioke Stephen Obioha v Public Prosecutor [2017] 1 SLR 1 at [14] and Daniel De Costa Augustin v Attorney-General [2020] 2 SLR 621 (“Daniel De Costa”) at [8]).

246 The right to express one’s sexual identity, even in private, is evidently not an express constitutional right. Nor is it a right that can be found in the Constitution, either as a matter of construing the Constitution in its entirety or as a matter of necessary implication (see Daniel De Costa at [9]). It is therefore impermissible to construe the unenumerated right to express one’s sexual identity as a right that attracts constitutional protection under Art 9(1).

247 Protection under Art 9(1) cannot extend to the protection of sexual identity for another reason – the words “life or personal liberty” in Art 9(1) refer only to freedom from unlawful deprivation of life and unlawful detention or incarceration. Such a restrictive reading of Art 9(1) is supported by the text, structure and history of the provision.

248 We first analyse the text and structure of Art 9(1). Article 9(1) and the other five sub-Articles in Art 9 all provide procedural safeguards in respect of the arrest and detention of persons. Article 9(2) refers to the common law prerogative writ of habeas corpus; Art 9(3) provides for an arrested person’s right to counsel and right to be informed of the grounds of his arrest; and Art 9(4) requires that an arrested person be produced before a magistrate “without unreasonable delay” and, in any case, within 48 hours. Articles 9(5) and 9(6) contain exceptions to the rights guaranteed to an arrested person. We echo the holding in Lim Meng Suang (CA) (at [46]) that, reading Art 9 as a whole, “the phrase ‘life or personal liberty’ in Art 9(1) refers only to a person’s freedom from an unlawful deprivation of life and unlawful detention or incarceration”.

249 This interpretation of “life or personal liberty” is confirmed by the history of Art 9(1). Article 9 is derived from Art 5(1) of the Constitution of the Federation of Malaya (“the 1957 Malayan Constitution”). The 1957 Malayan Constitution laid the foundation for the Constitution of Malaysia that came into effect when Malaysia (comprising the Federation of Malaya, Singapore, Sabah and Sarawak) was formed on 16 September 1963 (“the 1963 Malaysian Constitution”), and Art 5(1) of the latter was subsequently adopted in Singapore as Art 9(1) (see also Lim Meng Suang (CA) at [47] and Yong Vui Kong (MDP) at [61]–[63]).

250 The 1957 Malayan Constitution was drafted with advice from the Federation of Malaya Constitutional Commission chaired by Lord Reid (“the Reid Commission”), which had been appointed to make recommendations for a constitution for an independent Federation of Malaya (see Report of the Federation of Malaya Constitutional Commission 1957 (11 February 1957) (“the Reid Commission Report”) at para 1). The Reid Commission stated that Part II of the draft constitution, which was headed “Fundamental Liberties”, was intended to “define and guarantee certain fundamental individual rights which are generally regarded as essential conditions for a free and democratic way of life” (see the Reid Commission Report at para 161). It went on to elucidate the meaning of “personal liberty” as follows (see the Reid Commission Report at para 162):

It is clear from that extract that the Reid Commission regarded “personal liberty” in “any of its aspects” as relating to freedom from physical limitations to the person, which might take the form of detention without lawful authority, slavery, forced labour, banishment or restrictions on the freedom of movement.

251 Art 5(1) of the 1957 Malayan Constitution was, in turn, based on Art 21 of the Constitution of India (“the Indian Constitution”). While Art 21 of the Indian Constitution provided that “[n]o person shall be deprived of his life or personal liberty except according to procedure established by law” [emphasis added], Art 5(1) of the 1957 Malayan Constitution stated that “[n]o person shall be deprived of his life or personal liberty save in accordance with law” [emphasis added]. There is, however, no indication that the Reid Commission intended for Art 5(1) to be accorded a wider interpretation than Art 21. Indeed, the fact that Arts 5(2) to 5(4) protected the procedural rights of persons who had been arrested and detained suggests that Art 5(1) was similarly limited and procedural in nature. It follows that Art 9(1), which is identical to Art 5(1), should also be construed narrowly.

252 Thus, the text, structure and history of Art 9(1) mandate that the words “life or personal liberty” be interpreted restrictively to refer to only freedom from unlawful deprivation of life and unlawful detention or incarceration. It follows that these words do not encompass the freedom to express one’s sexual identity, and therefore that s 377A does not deprive a person of “life or personal liberty” in a sense that is caught by Art 9(1). Given this conclusion, it is strictly unnecessary for us to consider the argument that s 377A is “absurd” such that it does not constitute a valid law for the purposes of Art 9(1). Nonetheless, we address this issue in the next section for completeness.

253 Under Art 2(1) of the Constitution, “law” is defined as including:

The term “written law” is similarly defined in Art 2(1) as “this Constitution and all Acts and Ordinances and subsidiary legislation for the time being in force in Singapore”.

254 Section 377A is prima facie “written law” and thus “law” as defined in Art 2(1) of the Constitution. However, Art 9(1) does not justify all legislation that deprives a person of his life or personal liberty (see Yong Vui Kong (MDP) at [16] and [75]). The words “in accordance with law” under Art 9(1) have been interpreted to go beyond formal validity (in the sense of a valid enactment by the Legislature) to incorporate the following requirements:

(a) A statute must comply with the fundamental rules of natural justice, which are procedural rights aimed at securing a fair trial (see Yong Vui Kong (Caning) at [64]).

(b) A statute cannot be colourable legislation, such as legislation directed at securing the conviction of particular individuals (see Yong Vui Kong (MDP) at [16]).

(c) A statute cannot be absurd or arbitrary (see Yong Vui Kong (MDP) at [16]).

(d) A statute cannot be contrary to the rule of law (see Prabagaran a/l Srivijayan v Public Prosecutor and other matters [2017] 1 SLR 173 (“Prabagaran”) at [96]–[99]).

255 Before examining the parties’ arguments on whether s 377A is “absurd” such that it is not a valid law, it is necessary to first examine what the concept of “absurdity” entails.

256 To recapitulate, Mr Thuraisingam’s and Mr Ravi’s arguments on “absurdity” rest on the notion that s 377A criminalises individuals on the basis of an immutable facet of their identity. These arguments necessarily assume that the test of “absurdity” has substantive content. Mr Thuraisingam made this point explicitly by stating as follows:

257 Although Mr Thuraisingam does not set out an exhaustive list of “absurd” legislation, he contends that “one category of self-evidently ‘absurd’ legislation, is legislation that criminalises a class of persons for their identity, for no compelling interest” [emphasis in original in italics; emphasis added in bold italics]. This is, in our view, a substantive test predicated on a moral or value judgment.

258 On the other hand, Ms Tan argues that the test of “absurdity” only protects rights to a fair process – in other words, procedural rights. She submits that the test of “absurdity” should not and does not protect substantive rights for the following reasons:

(a) First, the Court of Appeal held in Yong Vui Kong (MDP) (at [80]) that the requirement that a law must not be “absurd” does not include or imply a requirement that it must be “fair, just and reasonable”. Such a requirement would require the court to intrude into the legislative sphere and engage in policy-making, which the court should refrain from doing, particularly as it is ill-equipped to make value judgments on polycentric matters.

(b) Second, Mr Thuraisingam and Mr Ravi have not provided any viable standard by which the court can judge a law to be so absurd as to be unconstitutional.

259 We agree with Ms Tan that the test of “absurdity” is procedural in nature and is intended to secure the right to a fair process in the context of a possible deprivation of life or personal liberty. An example of a statute that would be “absurd” by this yardstick is one that cannot be understood or complied with. It would be impossible to make a defence to any charge under such a statute, and it would thus be “absurd” to deprive a person of life or personal liberty on this basis. This understanding of “absurdity” also aligns with how our courts have interpreted the words “life or personal liberty” in Art 9(1) (see [247] and [252] above).

260 In Singapore, the notion that the “absurdity” of a statutory provision may arise for consideration in the context of Art 9(1) can be traced to the words of Lord Diplock at the hearing of Ong Ah Chuan v Public Prosecutor [1981] 1 AC 648 (“Ong Ah Chuan”). His Lordship specifically asked (at 659) the prosecutor whether he was contending that, “provided a statute is an Act of the Singapore Parliament[,] then however unfair or absurd or oppressive it may be[,] it is justified by article 9(1) of the Constitution” [emphasis added]. When the prosecutor answered that he was not advancing such an argument, and that it was in any event unnecessary for him to do so, Lord Diplock replied (likewise at 659): “Their Lordships cannot accept that because they will have to deal with the point. They are not disposed to find that article 9(1) justifies all legislation whatever its nature.” However, their Lordships did not elaborate in their judgment on the type of “absurd” legislation that might fall afoul of Art 9(1).

261 The concept of “absurdity” was revisited by our courts in Yong Vui Kong (MDP). Crucially, this concept was considered in the context of our seeking to explain what the Privy Council in Ong Ah Chuan had in mind during the oral exchange cited above. We suggested in obiter (at [16]) that:

262 Relying on the dicta emphasised in the above passage, Mr Thuraisingam argues that an “absurd” law refers to legislation that is so abhorrent that no reasonable person can contemplate it as being morally justified (see [256] above). We do not think that the aforesaid dicta in Yong Vui Kong (MDP) supports Mr Thuraisingam’s submission.

263 First, we reiterate that the aforesaid dicta in Yong Vui Kong (MDP) was obiter and was issued in the context of our attempt to explain what the Privy Council had meant during an oral exchange with the prosecutor at the hearing of Ong Ah Chuan. Furthermore, the concept of “absurdity” was not explored in Yong Vui Kong (MDP) as it was not a live issue before us. Although that case concerned the issue of what amounted to “law” or being “in accordance with law” for the purposes of Art 9(1), we did not explain the concept of “absurdity” or indicate whether it had any substantive content. Instead, we focused on the appellant’s submission that a substantive test of whether an impugned law was “fair, just and reasonable” ought to be applied to determine if that law was constitutional under Art 9(1) (see Yong Vui Kong (MDP) at [78]–[79]).

264 Second, in Yong Vui Kong (MDP), this court expressly rejected (at [80]) a substantive “fair, just and reasonable” test as the criterion for assessing the constitutionality of an impugned law under Art 9(1). This was because such a test would trespass onto Parliament’s territory and impermissibly require the court to engage in policy-making. This court further observed (likewise at [80]) that even in Ong Ah Chuan, the Privy Council had only held that, for the purposes of Art 9(1), any law depriving a person of his life or personal liberty had to be consistent with the “fundamental principles of natural justice”. In other words, the Privy Council did not lay down a separate test of “absurdity” in its judgment in Ong Ah Chuan. There is thus no basis for Mr Thuraisingam to suggest that Yong Vui Kong (MDP) permits judicial scrutiny of the substantive content of an impugned law in an Art 9(1) constitutional challenge. As Ms Tan aptly points out, our courts have repeatedly clarified that “the fundamental rules of natural justice referred to by the Privy Council in Ong Ah Chuan … [a]re ‘procedural rights aimed at securing a fair trial’” [emphasis added] (see Kho Jabing v Public Prosecutor [2016] 3 SLR 135 at [109] and Yong Vui Kong (Caning) at [62]–[64]).

265 For the reasons set out above, we are satisfied that the test of “absurdity” is procedural in nature and does not permit the court to examine the substantive content of s 377A. It follows that s 377A is self-evidently not “absurd”. Neither Mr Ravi nor Mr Thuraisingam has suggested that s 377A infringes any procedural rights, which are rights meant to ensure a fair trial.

266 Be that as it may, we engage with Dr Tan’s and Mr Ong’s case at its highest. For the sake of argument, we adopt Mr Thuraisingam’s test of “absurdity” – namely, that an impugned law is not “law” if it is so abhorrent that no reasonable person can contemplate it as being morally justified (see above at [256]). Even so, we do not agree that s 377A is “absurd”.

267 First, even if we were to accept that sexual orientation is immutable, this alone does not render s 377A “absurd” such that it is not a valid law. As we explained at [159] above, the argument that the Government can never regulate against immutable characteristics is clearly unsustainable.

268 Second, many reasonable people do in fact see s 377A as being morally justified, as is evident from the s 377A Debates. Numerous parliamentarians spoke up in favour of retaining s 377A, often on the ground of safeguarding societal morality and with the recognition that a sizable segment of our society regards homosexual behaviour as unacceptable. We reproduce three excerpts from the s 377A Debates by way of examples:

These views can hardly be dismissed as absurd by any reasonable measure. Rather, it might be said that these views and the views of the present appellants reflect competing notions of what is thought to be right and good. For the reasons outlined above, including those mentioned at [4]–[7], the courts should be wary of choosing between them, save to the extent that this follows from the application of specific legal principles.

269 To sum up, s 377A does not deprive a person of his personal liberty in a sense that is caught by Art 9(1). In any case, s 377A cannot be said to be so “absurd” as to fail to constitute a valid law for the purposes of Art 9(1). For these reasons, we agree with the Judge that s 377A does not violate Art 9(1).

270 In our judgment, the constitutional challenge under Art 14 by Dr Tan and Mr Choong is equally unmeritorious because the constitutional protection afforded under Art 14(1)(a) does not extend to acts of sexual intimacy. Again, we emphasise that our observations on this are strictly obiter since a finding on whether s 377A is constitutional under Art 14(1)(a) is unnecessary for the determination of these appeals.

271 Article 14 provides for, among other rights, the right to freedom of speech and expression. It is apposite to set out Art 14 in full:

272 It may be recalled that the Judge did not think that the right to freedom of expression could be divorced from the right to freedom of speech (see the Judgment at [249]; see also [38] above). He thus held that “the term ‘expression’ must be understood in its ordinary meaning to relate to freedom of speech encompassing matters of verbal communication of an idea, opinion or belief” (see the Judgment at [255]).

273 Mr Singh and Mr Ravi advance two main arguments in relation to Art 14(1)(a).

274 First, focusing on the proper construction of the words “freedom of … expression” in Art 14(1)(a), they argue that the Judge erred in interpreting those words in an unnecessarily restrictive manner:

(a) Mr Singh argues that, on a plain reading of the words “freedom of speech and expression”, the term “expression” should not be seen as surplusage. He thus submits that the Judge erred in confining the meaning of the term “expression” to “some form of verbal communication” (see the Judgment at [249]). He further contends that Art 14(1)(a) should be interpreted generously such that acts of sexual intimacy, which are “a fundamentally important form of personal expression”, constitute a form of protected “expression” under Art 14(1)(a). In Mr Singh’s words, consensual acts of sexual intimacy convey meaning as they are “the most intimate expressions of love and personal closeness”.

(b) Mr Ravi aligns himself with Mr Singh’s position. He rejects the Judge’s findings that the word “expression” in Art 14(1)(a) is surplusage and that “expression” refers to verbal expression only. In his view, an interpretation of Art 14(1)(a) that avoids construing the term “expression” as surplusage should be preferred. He further submits that Parliament must have intended to protect non-verbal modes of expression such as consensual sex, whether between homosexual or heterosexual individuals.

275 Second, Mr Ravi argues that s 377A infringes the right to freedom of speech by creating a chilling effect on gay rights advocacy.

276 As we see it, the core questions that must be answered are how the word “expression” in Art 14(1)(a) should be interpreted and, accordingly, whether the constitutional protection afforded under Art 14(1)(a) extends to acts of sexual intimacy. We analyse Mr Singh’s and Mr Ravi’s arguments according to the following sub-issues:

(a) What is the width and scope of the protection afforded under Art 14(1)(a)?

(b) Does s 377A infringe the right to freedom of speech and expression?

277 In considering the width and scope of the protection afforded under Art 14(1)(a), the central inquiry concerns the proper interpretation of the word “expression” as used in that provision, an analysis which once again calls for the application of the Tan Cheng Bock framework (see [162] above).

278 As the Judge recognised, a plain reading of the words “freedom of speech and expression” can give rise to a wide range of interpretations (see the Judgment at [243]–[244]). The ordinary meaning of the term “speech” is any form of communication expressed in the form of words, whether spoken or written. The dispute here centres on the term “expression”, which is usually regarded as being wider in scope than the term “speech”. The term “expression” can encompass any means of conveying meaning, opinions, beliefs or ideas, whether with or without the use of language. We agree with the Judge that the plain and ordinary meaning of the word “expression” in the abstract does not rule out the possibility of sexual intercourse being a form of expression (see the Judgment at [244]).

279 However, having regard to the context of Art 14(1)(a) in the Constitution, it is clear to us that the primary right protected thereunder is that of “freedom of speech” and not “freedom of expression”. This can be gleaned from the marginal note to Art 14. The marginal note omits any mention of “freedom of expression” and focuses solely on “freedom of speech”. Even though the meaning of a statutory provision must ultimately be gleaned from the statutory language as well as the context of the provision, it is well established that marginal notes can be used as an aid to statutory interpretation (see Tee Soon Kay v Attorney-General [2007] 3 SLR(R) 133 at [41]). While a marginal note is not meant to provide a comprehensive and fully accurate summary of the provisions that it covers, it provides a brief indication of the content of those provisions (see Diggory Bailey & Luke Norbury, Bennion on Statutory Interpretation (LexisNexis, 7th Ed, 2017) at section 16.7) and forms part of the relevant context within which those provisions should be interpreted.

280 The marginal note to Art 14 reads: “Freedom of speech, assembly and association”. The constitutional freedoms reflected in the marginal note correspond, in order, to the three sub-clauses of Art 14(1). The marginal note indicates the main subject matter of each of these sub-clauses, namely: speech, assembly and association. It can also be observed from the wording of Arts 14(1)(a), 14(1)(b) and 14(1)(c) that these sub-clauses qualify, restrict and describe the nature of the broad rights listed in the marginal note. This is, in itself, a pattern of statutory interpretation. To be more precise, what is at play is the noscitur a sociis principle of construction, which provides that “words, and particularly general words, cannot be read in isolation: their colour and content are derived from their context” (see Public Prosecutor v Lam Leng Hung and others [2018] 1 SLR 659 (“Lam Leng Hung”) at [107]–[108], citing the House of Lords’ decision in Attorney-General v Prince Ernest Augustus of Hanover [1957] AC 436 at 461).

281 By way of illustration, even though Art 14(1)(b) sets out “the right to assemble peaceably and without arms”, the marginal note simply refers to “[f]reedom of … assembly”. The core constitutional right protected under Art 14(1)(b) is thus the right to freedom of assembly; the phrase “peaceably and without arms”, as used in Art 14(1)(b), qualifies the right to freedom of assembly and describes the nature of that right. Similarly, Art 14(1)(c) qualifies and describes the nature of the constitutional right to freedom of association by clarifying that what is protected is the right to “form associations” [emphasis added].

282 While the term “expression” is ordinarily seen as being broader in scope than the term “speech”, there is no mention of the right to freedom of expression as a free-standing right in the marginal note to Art 14. This suggests that the term “expression” was not meant to unilaterally expand the scope of Art 14(1)(a) beyond the right to freedom of speech. On the contrary, applying the noscitur a sociis principle to Art 14(1)(a), the term “expression” assumes the ancillary function of qualifying and describing the fundamental right enshrined in Art 14(1)(a) – which is, as the marginal note to Art 14 indicates, the right to freedom of speech. Although this means that the term “expression” as used in Art 14(1)(a) has a more restrictive meaning than its literal or usual meaning, this conclusion is hardly aberrant and is simply the result of applying the well-established principle of noscitur a sociis (see Lam Leng Hung at [110]).

283 Having similarly observed that the marginal note to Art 14 makes no mention of the right to freedom of expression as a free-standing right, the Judge concluded that that right was “contemplated as something relating to or falling within the right to freedom of speech” (see the Judgment at [246]). We prefer to see the word “expression” as giving colour to the term “speech” (see Lam Leng Hung at [109]).

284 The upshot of our analysis is that Art 14(1)(a) protects the right to freedom of speech – that is to say, any form of communication that is expressed in words, whether spoken or written – that conveys meaning, opinions, beliefs or ideas. Since acts of sexual intimacy are not “speech” to begin with, the Art 14 constitutional challenge to s 377A is without merit.

285 Aside from the marginal note to Art 14, there is another reason why the term “expression” in Art 14(1)(a) must be interpreted restrictively (as stated at [282] above) and not in its usual broad sense (namely, any means of conveying meaning, opinions, beliefs or ideas, whether with or without the use of language: see [278] above). An expansive reading of the term “expression” would make it extremely difficult to delimit the acts that are constitutionally protected under Art 14(1)(a). Virtually any act that purports to convey meaning, opinions, beliefs or ideas, even an act amounting to a sexual offence (such as necrophilia or bestiality), could claim protection under Art 14(1)(a) as a form of “expression”. Such a plainly absurd result would be contrary to the well-established canon of statutory interpretation that Parliament does not legislate with the intention of producing unworkable or impracticable results (see Tan Cheng Bock at [38]).

286 In response to this, Mr Ravi argues that heinous acts such as necrophilia or bestiality would justify derogation from the right to freedom of speech and expression, ostensibly on the ground of public morality under Art 14(2)(a). Accepting this for the sake of argument, it should nonetheless be noted, however, that Art 14(2)(a) lists the grounds on which Parliament may restrict “the rights conferred by [Art 14(1)(a)]” [emphasis added]. The absurdity thus lies in the recognition that acts such as necrophilia and bestiality do fall within “the rights conferred by [Art 14(1)(a)]”, albeit that those rights may be lawfully curtailed in this context. As Parliament must be presumed not to have intended such an absurd result, there is good reason to prefer a more restrictive interpretation of the term “expression”, as used in Art 14(1)(a), than what Mr Singh and Mr Ravi argue for.

287 Mr Ravi further contends that the potentially untrammelled scope of the right to freedom of speech and expression can be reined in by adopting the following test: acts of non-violent expression are only protected under Art 14(1)(a) in so far as they are “performed to convey a meaning” [emphasis in original omitted], and whether a particular act fulfils this description is a question to be decided on the facts of each case. In proposing this test, he draws inspiration from the majority judgment in the Supreme Court of Canada case of Irwin Toy Ltd v Québec (Attorney General) [1989] 1 SCR 927 (“Irwin Toy”).

288 Mr Ravi’s proposed solution, however, does little to reduce the potential expansiveness of the right to freedom of speech and expression that would follow from his interpretation of Art 14(1)(a). In Irwin Toy, the majority of the court held (at [53]) that:

289 The problem is that almost any act can be said to convey some kind of meaning and to relate to the pursuit of truth, participation in the community, individual self-fulfilment or human flourishing. We therefore reiterate that the term “expression” in Art 14(1)(a) should be interpreted as a qualifier of the right to freedom of speech. Consequently, Art 14(1)(a) protects the right to freedom of speech – that is to say, any form of communication that is expressed in words, whether spoken or written – that conveys meaning, opinions, beliefs or ideas (see [284] above).

290 Having interpreted the term “expression” in Art 14(1)(a) at the first stage of the Tan Cheng Bock framework, we now turn to the second stage of that framework, which calls for the legislative purpose of Art 14(1)(a) to be discerned. For the reasons set out at [279]–[289] above, we consider that the legislative purpose of Art 14(1)(a) is the protection of free speech. This is also confirmed by the extraneous material.

291 Of relevance in this regard is the 1966 Constitutional Commission Report, which we referred to at [39] above. This report was prepared by the Constitutional Commission chaired by then Chief Justice Wee Chong Jin. The Constitutional Commission was appointed to look into, among other things, the protection of minority rights in Singapore and the provisions that ought to be entrenched in the then version of the Constitution after Singapore had become a sovereign republic. This is an important document concerning the fundamental liberties presently protected under Part IV of the Constitution because several recommendations contained therein were accepted by the Government. Paragraph 26 of the 1966 Constitutional Commission Report stated that the Constitutional Commission was:

292 In respect of Art 14, the 1966 Constitutional Commission Report stated (at para 37) as follows:

The abovementioned Art 10 of the 1963 Malaysian Constitution subsequently became the present-day Art 14.

293 We note that the italicised words in the above passage mirror the marginal note to Art 14(1)(a). This indicates that the marginal note was a considered and accurate reflection of the core right protected by each of the three sub-clauses of Art 14(1), thereby fortifying our interpretation of Art 14(1)(a).

294 To sum up, given our finding that the focus of Art 14(1)(a) is on the right to freedom of speech, the Art 14 constitutional challenge to s 377A must proceed from the premise that the term “speech” includes acts of gross indecency. However, such acts are not “speech” to begin with; in any event, extending the protection afforded under Art 14(1)(a) to acts of gross indecency would generate an absurd result that could not have been intended by the constitutional draftsmen. On these grounds, we consider that the Art 14 constitutional challenge must fail.

295 We turn to Mr Ravi’s argument that s 377A infringes the right to freedom of speech under Art 14(1)(a) by creating a chilling effect on gay rights advocacy (see [275] above). Mr Ravi contends that the criminalisation of sexual acts between men creates a climate that is hostile to the discussion of anything LGBT-related, thereby stymieing gay rights advocacy. In our judgment, Mr Ravi’s argument must be rejected for three reasons.

296 First, the mere fact that a particular cause or activity is presently illegal does not in and of itself generate a chilling effect that stifles advocacy of the prohibited cause or activity. Gay rights activism and advocacy continue to exist in Singapore in a variety of forms, an example being the highly publicised Pink Dot Singapore event that is held annually.

297 Second, while gay rights advocates may face challenges in their work, it is likely that their difficulties stem in large part from general societal attitudes, as opposed to the continued retention of s 377A in particular. In other words, any chilling effect on gay rights advocacy that may currently exist has not been shown to be attributable to s 377A specifically.

298 Third, the claim that s 377A has a chilling effect on gay rights advocacy is really an extra-legal argument made under the guise of an alleged violation of Art 14(1)(a). What Mr Ravi is essentially arguing is that the striking down of s 377A would promote gay rights advocacy. Such an argument goes towards the socio-political desirability of retaining or repealing s 377A, which is not a matter that can be properly considered by the court.

299 In summary, s 377A does not engage the right to freedom of speech and expression protected under Art 14(1)(a). We therefore see no merit in the Art 14 constitutional challenge.

300 Finally, we turn to Mr Ravi’s and Mr Singh’s contention that s 377A fails the “reasonable classification” test and thus falls afoul of Art 12. In this section, we will also address Mr Thuraisingam’s argument that s 377A is “arbitrary” and hence in violation of Art 9(1) (see [241] above). We stress once again that our observations below are purely obiter.

301 We begin our analysis by examining the structure of Art 12. Articles 12(1) and 12(2) provide as follows:

302 Article 12(1) provides for equality before the law and the equal protection of the law for all persons. As noted in Lim Meng Suang (CA) at [90], Art 12(1) is framed at a very general level and is in the nature of a declaratory statement of principles.

303 The rights enshrined in Art 12(1) are so fundamental and basic that the court should eschew any approach that renders Art 12(1) toothless. This means that while Parliament has a wide ambit to legislate, there is no presumption that every differentiating measure that it enacts bears a rational relation to the object sought to be achieved. Such a presumption of constitutionality under Art 12 would be impermissible because “relying on a presumption of constitutionality to meet an objection of unconstitutionality would entail presuming the very issue which is being challenged” (see Saravanan at [154]).

304 In contrast to Art 12(1), Art 12(2) is less open-ended. This is unsurprising given that Art 12(2) specifies the exclusive grounds on which discrimination against Singaporean citizens is absolutely barred, namely: religion, race, descent and place of birth. The grounds specified in Art 12(2) indicate the types of diversities that have been deemed worthy of greater constitutional protection, thereby reflecting the ethno-religious, pluralistic community that the Constitution is designed to sustain: see Thio Li-Ann, A Treatise on Singapore Constitutional Law (Academy Publishing, 2012) at para 13.009.

305 The established test for assessing whether a statutory provision is constitutional under Art 12 is the “reasonable classification” test (see Lim Meng Suang (CA) at [60] and Public Prosecutor v Taw Cheng Kong [1998] 2 SLR(R) 489 (“Taw Cheng Kong (CA)”) at [54] and [58]). Under this test, a statutory provision which prescribes a differentiating measure will be consistent with Art 12(1) only if: (a) the classification prescribed by the provision is founded on an intelligible differentia; and (b) that differentia bears a rational relation to the object sought to be achieved by the provision. We refer to these as the first and second limbs of the “reasonable classification” test respectively. The “reasonable classification” test is also the applicable test for determining whether a law is “arbitrary” and hence in contravention of Art 9(1) (see Prabagaran at [93]).

306 The “reasonable classification” test rests on the notion that like cases should, in broad terms, be treated alike. The underlying rationale behind this test was explained in Saravanan as follows (at [153]):

307 However, the fundamental rubric that “like cases should, in broad terms, be treated alike” does not inform us of the level of abstraction at which individuals should be grouped into classes so that the legitimacy of the differential treatment in question may be properly assessed (see Lim Meng Suang and another v Attorney-General [2013] 3 SLR 118 (“Lim Meng Suang (HC)”) at [60]). In other words, the maxim that “like cases should be treated alike” alone cannot guide the practical application of the “reasonable classification” test. For this reason, we turn to examine how this test has been applied in the case law, focusing on the approaches adopted in Lim Meng Suang (CA) and, more recently, Syed Suhail bin Syed Zin v Attorney-General [2021] 1 SLR 809 (“Syed Suhail”).

308 The nature and function of the “reasonable classification” test were considered at length in Lim Meng Suang (CA), in which the court characterised the test as one that serves the “minimal threshold function of requiring logic and coherence in the [statutory provision] concerned” [emphasis in original] (at [66]).

309 It will be recalled that the first limb of the “reasonable classification” test requires that the classification prescribed by the statutory provision in question be based on an intelligible differentia. The court emphasised in Lim Meng Suang (CA) (at [65]) that this requirement was “a relatively low threshold that ought to avoid any consideration of substantive moral, political and/or ethical issues because these issues are potentially (and in most instances, actually) controversial” [emphasis in original]. The court also acknowledged that on this understanding of the “reasonable classification” test, a statutory provision would “very seldom” [emphasis in original] fail to pass muster under the first limb of the test, which set out only a “threshold legal criterion” [emphasis in original]. Citing Lim Meng Suang (HC) at [47], the court held in Lim Meng Suang (CA) (at [65]) that the first limb of the “reasonable classification” test calls for nothing more than that the relevant differentia “be understood or [be] capable of being apprehended by the intellect or understanding, as opposed to by the senses”.

310 However, the court caveated in Lim Meng Suang (CA) (at [67]) that a differentia which is capable of being understood or apprehended by the intellect or understanding may nevertheless “still be unintelligible to the extent that it is so unreasonable as to be illogical and/or incoherent” [emphasis in original]. In this regard, the illogicality and/or incoherence must be “so extreme that no reasonable person would ever contemplate the differentia concerned as being functional as [an] intelligible differentia” [emphasis in original].

311 As for the second limb of the “reasonable classification” test, it was noted in Lim Meng Suang (CA) (at [68]) that what this limb required was “a rational relation” [emphasis in original] between the differentia embodied in and the legislative object of the statutory provision in question. Further, “the requisite rational relation will – more often than not – be found”, not least because a perfect relation or complete coincidence is not required (see Lim Meng Suang (CA) at [68]). That said, the court ultimately held in Lim Meng Suang (CA) (at [153]) that there was “a complete coincidence” [emphasis in original] between the differentia embodied in and the legislative object of s 377A. While the court did not explicitly articulate the legislative object of s 377A, it appeared to have adopted the first instance judge’s finding that s 377A was intended to criminalise male homosexual conduct owing to the perceived undesirability of such conduct (see Lim Meng Suang (CA) at [24] and [27]). Given how the legislative object of s 377A was framed in Lim Meng Suang (CA), it is unsurprising that the court went on to find a complete coincidence between that legislative object and the differentia embodied in s 377A.

312 In Lim Meng Suang (CA), the court also considered (at [114]) the hypothetical example of a law banning all women from driving. The court suggested that it would be at least arguable that such a law would not pass muster under the “reasonable classification” test, chiefly because the differentia embodied in that law might be illogical and/or incoherent under the first limb of the test. Moreover, the court thought it to be at least arguable that there might not be a rational relation between the differentia embodied in and the object sought to be achieved by such a law – unless the object of the law was precisely to ban all women from driving. The court indicated, however, that if that were the case, one would have to return to the question whether the differentia embodied in such a law was illogical and/or incoherent under the first limb of the “reasonable classification” test. Therefore, under the Lim Meng Suang (CA) approach, such a law would arguably fail both limbs of the test.

313 Since the decision in Lim Meng Suang (CA), another approach to the application of the “reasonable classification” test has emerged in our decision in Syed Suhail. This was a case concerning administrative review of an executive action, rather than constitutional review of a statutory provision. As will be discussed below, there are two main differences between the Lim Meng Suang (CA) and the Syed Suhail approaches, namely:

(a) whether, albeit in an extreme minority of cases, as suggested in Lim Meng Suang (CA), the first limb of the “reasonable classification” test permits consideration of the reasonableness (or lack thereof) of the differentia embodied in the statutory provision in question; and

(b) the level of scrutiny to which the impugned statutory provision is subject, having regard to its correctly identified object and its relationship with the differentia employed.

314 Before delving into these differences, we consider first the commonality in the two approaches. This lies in the first limb of the “reasonable classification” test, in so far as it serves the purpose of ensuring that there is a differentia that is capable of being assessed for legality under the second limb of the test. In Lim Meng Suang (CA), we explained (at [64]) why the first limb of the test necessarily operates prior to the second limb: if there is no intelligible differentia to begin with, then there is no intelligible differentia that can be assessed against the legislative object of the statutory provision in question. Similarly, in Syed Suhail, we described (at [62]) the differential criterion as “an analytical tool used to isolate the purported rationale for differential treatment, so that its legitimacy may then be assessed properly” [emphasis added]. The Lim Meng Suang (CA) and the Syed Suhail approaches are thus aligned to the extent that, under both approaches, the first limb of the “reasonable classification” test is concerned with ensuring that the differentia embodied in the statutory provision in question is one that can be assessed against the legislative object of that provision. This objective of the first limb of the test is realised under the Lim Meng Suang (CA) approach by requiring that the differentia in question be capable of being understood or apprehended by the intellect or understanding; and under the Syed Suhail approach, by identifying the purported criterion for the differential treatment in question.

315 The first main difference between the Lim Meng Suang (CA) and the Syed Suhail approaches lies in whether, in extreme cases, the first limb of the “reasonable classification” test permits consideration of the reasonableness (or lack thereof) of the differentia embodied in the statutory provision concerned. Under the Lim Meng Suang (CA) approach, the validity of a differentia may be called into question under the first limb of the “reasonable classification” test in extreme cases where the differentia is “so unreasonable as to be illogical and/or incoherent” [emphasis in original] (at [67]). Although this line from Lim Meng Suang (CA) contains the words “illogical” and “incoherent”, it is evident that the court was ultimately concerned with the reasonableness (or lack thereof) of a differentia, albeit only in extreme cases.

316 In this regard, we refer to the hypothetical example of a law banning all women from driving, which was raised in Lim Meng Suang (CA) at [114] and which we referred to at [312] above. It will be recalled that the court in Lim Meng Suang (CA) thought it arguable that such a law might fail the first limb of the “reasonable classification” test because the differentia embodied in that law might be “so unreasonable as to be illogical and/or incoherent” [emphasis in original] (at [67]). We return to this example at [319] below.

317 It therefore seems to us that, under the Lim Meng Suang (CA) approach, whether a differentia is “so unreasonable as to be illogical and/or incoherent” [emphasis in original omitted] goes beyond whether that differentia is capable of being understood or apprehended by the intellect or understanding. After all, a differentia that is so unreasonable that it fails the first limb of the “reasonable classification” test may nonetheless be capable of being so understood or apprehended. Instead, in so far as such extreme cases are concerned, the question whether a differentia is “so unreasonable as to be illogical and/or incoherent” [emphasis in original omitted] inherently entails a judgment on the reasonableness (or lack thereof) of that differentia.

318 In contrast, under the Syed Suhail approach, the first limb of the “reasonable classification” test is never concerned with the reasonableness (or lack thereof) of the differentia in question, even in cases where the differentia is extremely unreasonable. As we mentioned at [314] above, under the Syed Suhail approach, the first limb of the “reasonable classification” test is only concerned with identifying the purported criterion for the differential treatment in question. It is only at the second limb of the test that the court considers “whether the differential treatment [is] reasonable” [emphasis added] (see Syed Suhail at [61]). In the context of a challenge to the constitutionality of a statutory provision under Art 12, the relevant inquiry under the second limb is whether the differential treatment bears “a sufficient rational relation to … the object of the statutory provision” (see likewise Syed Suhail at [61]). Hence, under the Syed Suhail approach, a grossly unreasonable differentia of the kind contemplated in Lim Meng Suang (CA) at [67] would fail the second but not the first limb of the “reasonable classification” test.

319 Here, we return to the hypothetical example of a law banning all women from driving (see [312] and [316] above). As we mentioned earlier, it was suggested in Lim Meng Suang (CA) at [114] that such a law might fail the first limb of the “reasonable classification” test as its differentia could plausibly be said to be so patently illogical and/or incoherent that “no reasonable person would ever contemplate [it] as being functional as [an] intelligible differentia” [emphasis in original] (see Lim Meng Suang (CA) at [67]). However, one would likely reach a different conclusion if one were to apply the Syed Suhail approach. Under this approach, the first limb of the “reasonable classification” test is, in all cases, concerned only with identifying the purported criterion for the differential treatment in question. In so far as this hypothetical example is concerned, the purported criterion for differential treatment is readily ascertainable – namely, gender. Hence, under the Syed Suhail approach, a law banning all women from driving would likely fail the “reasonable classification” test not because the gender-based differentia embodied in that law is so illogical and/or incoherent that it fails the first limb of the test, but because the differentia bears no rational relation to any conceivable object of that law under the second limb of the test. In other words, under the Syed Suhail approach, such a law would fail the second limb of the “reasonable classification” test, even though it would be found to have an intelligible differentia under the first limb of the test.

320 As against this, it might be argued that a law banning all women from driving might not necessarily fail the second limb of the “reasonable classification” test under the Syed Suhail approach. In this regard, it might be said that if the object of that law is precisely to ban all women from driving, there would be a complete coincidence between the gender-based differentia embodied in and the object of that law. The rebuttal to this argument lies in framing the object of a law that is challenged under Art 12 at the appropriate level of generality, a point that we discuss in greater detail at [322]–[324] below. For now, it suffices for us to highlight that framing a ban on all women from driving as the very object of a law would be tantamount to saying that the object of that law is to introduce the differentia that it embodies – which is circular in reasoning. Put another way, the level of generality at which the object of a law is pitched would effectively be determined by the differentia embodied in that law, and there would necessarily be a perfect relation between the differentia and the object of that law. Given that the Syed Suhail approach is averse to the application of the “reasonable classification” test in any manner that renders it purely formalistic (see [325]–[327] below), it would be impermissible, under that approach, to frame the object of a law banning all women from driving as precisely that – to ban all women from driving.

321 To reiterate the first key difference between the Lim Meng Suang (CA) and the Syed Suhail approaches, a differentia that is patently unreasonable would arguably fail both limbs of the “reasonable classification” test under the Lim Meng Suang (CA) approach; but, applying the Syed Suhail approach, such a differentia would likely fail the second limb of the test. If the identification of the differentia embodied in a statutory provision is simply meant to “isolate the purported rationale [or criterion] for differential treatment” (see Syed Suhail at [62]), then the court should not read more into the first limb of the “reasonable classification” test. It should be noted, however, that to the limited extent that the Lim Meng Suang (CA) approach incorporates (under the first limb of the “reasonable classification” test) a substantive evaluation of the reasonableness of the differentia in question, that approach overlaps substantially with the Syed Suhail approach (albeit under the second limb of the test). Indeed, in the example considered at [312], [316] and [319] above, both approaches result in the same conclusion that a law banning all women from driving would violate Art 12.

322 To appreciate the second area of difference between the Lim Meng Suang (CA) and the Syed Suhail approaches, we must first consider the second limb of the “reasonable classification” test under the Syed Suhail approach in some detail. Since the second limb pertains to whether there is a rational relation between the differentia embodied in and the legislative object of an impugned statutory provision, a key issue that merits consideration is the level of specificity at which the legislative object ought to be pitched. The “paramount importance” of properly framing the legislative object of a statutory provision was underscored in Tan Cheng Bock at [39]. Referring to the minority judgment in Attorney-General v Ting Choon Meng and another appeal [2017] 1 SLR 373 at [60], we highlighted (likewise at [39]) that “[c]asting the legislative purpose differently or at different levels of generality may result in varying and even conflicting interpretations”. If one were to articulate the legislative object of a statutory provision in whatever terms would support one’s desired interpretation of the provision, the “reasonable classification” test could be reduced to nothing more than an exercise in legal formalism.

323 The importance of pitching the legislative object of a statutory provision at the appropriate level of generality is evident from the contrasting decisions reached by the High Court and this court in Taw Cheng Kong v Public Prosecutor [1998] 1 SLR(R) 78 (“Taw Cheng Kong (HC)”) and Taw Cheng Kong (CA) respectively (see the discussion of both cases in Lim Meng Suang (HC) at [52]–[60]). The statutory provision under consideration in both cases was s 37(1) of the Prevention of Corruption Act (Cap 241, 1993 Rev Ed) (“the PCA”), which provided that a Singapore citizen who committed an offence under the PCA in any place outside Singapore could be dealt with in respect of the offence as if it had been committed in Singapore. In Taw Cheng Kong (HC) at [51], the High Court construed the legislative object of the PCA narrowly as the eradication of corruption from the Singapore civil service and among fiduciaries in Singapore (see Taw Cheng Kong (HC) at [51]). This precipitated the High Court’s finding that the differentia of citizenship embodied in s 37(1) of the PCA did not bear a reasonable nexus to the legislative object of the PCA. In contrast, we framed the legislative object of the PCA much more broadly as the control and suppression of corruption, including extra-territorial corruption (see Taw Cheng Kong (CA) at [63]). On this basis, we went on to conclude that there was a rational relation between the differentia of citizenship in s 37(1) of the PCA and the legislative object of that statute.

324 The present appeals likewise illustrate how casting the legislative object of a statutory provision differently may well yield different results under the “reasonable classification” test. If one were to frame the legislative object of s 377A as the expression of societal disapproval of male-male sex acts, as Ms Tan asserts, there would necessarily be a perfect coincidence between the differentia embodied in and the legislative object of s 377A. On the other hand, if one were to cast the legislative object of s 377A more broadly as the expression of societal disapproval of homosexual conduct in general or the safeguarding of public morality generally, that would strengthen the case that s 377A falls afoul of the “reasonable classification” test. In this setting, s 377A would appear to be under-inclusive because it does not criminalise female-female homosexual conduct, for instance. One could then conclude that the differentia embodied in s 377A (namely, male-male sex acts) lacks a rational relation to the legislative object of reflecting societal disapproval of homosexual conduct in general or safeguarding public morality generally. The framing of the legislative object of a statutory provision could, of course, cut both ways. For example, if Dr Tan and Mr Choong had been able to argue successfully in their respective appeals that the legislative object of s 377A is targeted specifically at male prostitution only, they might have succeeded in showing that as s 377A applies to categories outside the narrow category just mentioned (namely, male prostitution), it is over-inclusive and, hence, unconstitutional under Art 12.

325 It has been frequently noted that there need not be a “perfect relation” or a “complete coincidence” between the differentia embodied in and the legislative object of a statutory provision in order for the second limb of the “reasonable classification” test to be satisfied (see, for instance, Lim Meng Suang (CA) at [68], which we referred to at [311] above). But the Syed Suhail approach emphasises that while that relationship need not be perfect, it also must not be so tenuous as to be incapable of withstanding scrutiny. This is so for two reasons: (a) the nature of the rights at stake; and (b) the constitutional role of the court.

326 First, equality before the law and the equal protection of the law are such fundamental rights that the court should shun the “austerity of tabulated legalism” when applying the “reasonable classification” test, so that individuals can enjoy the “full measure” of their Art 12 rights (see Ong Ah Chuan at 670). Second, whether a statutory provision breaches Art 12 is a matter for the court’s determination. If the “reasonable classification” test were too diluted, that would undermine the court’s constitutional role in safeguarding individuals’ Art 12 rights. The court should therefore be chary of construing or applying the “reasonable classification” test in a manner that effectively denudes Art 12 of real force.

327 This brings us to the second key difference between the Lim Meng Suang (CA) and the Syed Suhail approaches. In Syed Suhail, this court recognised (at [63]) that when applying the “reasonable classification” test, the context can affect how stringently a statutory provision should be scrutinised. Syed Suhail concerned a prisoner who was facing imminent execution. He sought leave to commence judicial review proceedings against his impending execution based on, among other grounds, the scheduling of his execution before that of other prisoners who had been sentenced to death ahead of him. This court held (at [63]) that:

In contrast, the “reasonable classification” test was described in Lim Meng Suang (at [66]) as one that serves the “minimal threshold function of requiring logic and coherence in the [statutory provision] concerned” [emphasis in original].

328 In summary, the Lim Meng Suang (CA) and the Syed Suhail approaches to the “reasonable classification” test differ in two key respects. First, the first limb of the “reasonable classification” test assumes greater significance under the Lim Meng Suang (CA) approach, albeit only in cases where the differentia in question is so illogical and/or incoherent that no reasonable person would ever contemplate it as being capable of functioning as an intelligible differentia (at [67]). In contrast, under the Syed Suhail approach, the reasonableness (or lack thereof) of a differentia falls to be considered only under the second limb of the “reasonable classification” test. Second, the Syed Suhail approach contemplates a higher level of scrutiny when evaluating whether a statutory provision satisfies the “reasonable classification” test, particularly if the provision has a significant bearing on an individual’s life and liberty. In contrast, under the Lim Meng Suang (CA) approach, the “reasonable classification” test is of a “threshold nature” and is only meant to sift out laws which are “patently illogical and/or incoherent” (see Lim Meng Suang (CA) at [70]–[71]). For the avoidance of doubt, the Syed Suhail approach does not afford the court an open-ended mandate to evaluate legislation on the basis of its policy preferences, for that would be outside its constitutional role. This is something we have repeatedly and consistently emphasised. Indeed, and correlatively, there might even be a difference when considering statutory provisions as compared to executive action for compatibility with Art 12.

329 Given our finding at [149] above that s 377A is currently unenforceable in its entirety, it is unnecessary for us to decide whether the Lim Meng Suang (CA) approach or the Syed Suhail approach to the “reasonable classification” test should be preferred. We have set out the broad differences between these two approaches and consider that this issue merits further reflection on a suitable occasion in the future.

330 As we have repeatedly emphasised, the court cannot take a blinkered approach to the political reality surrounding s 377A and the implications of this context on the constitutionality of the provision. The political compromise on s 377A that was struck by the Government in 2007 was later echoed and elaborated on by AG Wong in 2018. AG Wong’s representations have engendered legitimate expectations which deserve legal protection so that the political compromise on s 377A may be properly upheld. We have held that to give full effect to the political compromise and, in turn, AG Wong’s representations in a legally acceptable manner, the entirety of s 377A is unenforceable unless and until the AG of the day provides clear notice that he, in his capacity as the PP: (a) intends to reassert his right to enforce s 377A proactively by way of prosecution; and (b) will no longer abide by the representations made by AG Wong in 2018 as to the prosecutorial policy that applies to conduct falling within the Subset. Having answered the Anterior Question in this way, it is unnecessary for us to address the constitutional questions raised by the appellants. They do not face any real and credible threat of prosecution under s 377A at this time and therefore do not have standing to pursue their constitutional challenges to that provision.

331 Given the unusual nature of these proceedings, the important questions of public interest that were raised and the findings that we have arrived at, we make no order as to costs for the proceedings both before us and in the court below. The usual consequential orders will apply.