1 It was in the evening of 30 September 2025 when the Applicant left the matrimonial home. She was carrying [A], her five-month-old son, in a baby carrier strapped to her chest. She did not take much with her. She went to her mother’s home and did not return. She filed a police report two weeks later. These are the bare facts of how this case came to court. Behind them lies a story that began, as many family violence cases do, not with a dramatic incident but with the slow accumulation of smaller ones. A newly forged marriage. A newborn child, only born months earlier in the April of 2025. An argument about the temperature of water used for formula milk preparation that ended not in resolution or compromise, but with a footrest kicked aside, a mobile phone snatched from the Applicant’s hand and thrown onto the sofa, slightly less than a month postpartum. Four months after that came the kick that ended the marriage’s first year and brought her, on that September night, to the door with her infant son and nowhere else to go. How should the law respond to conduct of this kind? The kind of conduct which leaves no bruises, that a man on the street might not recognise as family violence, but that a person living within it may experience as a slow intrusion into her sense of safety. This is the question at the heart of this judgment.

2 What conduct amounts to emotional or psychological abuse? The words are not strangers to everyday speech. If anything, they have never been more familiar. A generation that has grown up with the language of trauma, narcissism, toxicity and gaslighting, absorbed not only from therapists and self-help books but from social media feeds and online communities dedicated to naming and sharing the experience of harm, reaches for these concepts with a fluency that would have been unimaginable a decade ago. That growing awareness has given people the vocabulary to not only recognise but also articulate what was done to them. But vocabulary is not the law. The law demands more than a label. It asks a court to distinguish between the pain that is an inevitable feature of human relationships (eg, the unkind word, the door shut too firmly, the silence that lingers too long), and conduct that has crossed into something the law recognises as abuse, for which it will restrain, and if necessary, punish. Where exactly that line falls under the new provisions of the Women’s Charter 1961 (2020 Rev Ed) (“the Charter”) is a question that the courts are still working out.

3 The Applicant applied for a Personal Protection Order (“PPO”) against the Respondent pursuant to s 60A of the Charter. Both parties were unrepresented and presented their cases in person. I delivered my oral decision on 11 March 2026, granting the PPO and making a counselling order of the court’s own motion under s 60E(2)(a) read with s 60E(3)(a). This marriage is, by any measure, still young. It was not yet two years old at the date of this judgment and had been tested almost from its first months by the pressures of new parenthood. Counselling may assist them in developing the communication and conflict resolution strategies that [A]’s upbringing will require of them, whatever the future their new marriage holds. The Respondent has appealed my decision to the High Court.

4 These are my written grounds. So far as I am aware, this remains among the first cases arising under the new emotional or psychological abuse provisions of the Charter. The provisions have been in force for little over a year. The law is still developing and it is doing so against a backdrop of a growing number of applications to the Family Courts (a point explored at [14] below). The uncertainty that presently attends the scope of s 58B(4) and the standard that governs its application, has been a source of practical difficulty not only for potential applicants and their advisers but also for the courts themselves in navigating the new terrain. Even though I had earlier delivered my oral grounds to the parties at the conclusion of the hearing on 11 March 2026, it is with the novel issues engaged in the present case in mind that I provide these fuller grounds.

5 The reader who seeks a summary of the rather lengthy discussion on the scope of emotional or psychological abuse under s 58B(4) and the standard governing its application is directed to [103]–[107] below. These grounds are by the standards of a first instance decision lengthy. However, this reflects the genuine uncertainty that presently attends the new landmark amendments. As this may very well be the first case to be placed before the High Court for guidance, I have undertaken an extended analysis of the issues brought on by the new provisions not just to scope out the outstanding issues, but also to fully develop my reasons for how those issues may be addressed.

6 The parties were married sometime in July 2024. They have one child together, [A], born 11 April 2025. The first year of this marriage thus saw the birth of their son. [A] was approximately five months old at the time of the most recent incident. Following their marriage, the parties moved into the Respondent’s family home, a five-bedroom apartment, shared with the Respondent’s parents, his youngest brother, and a domestic helper. The Respondent’s other siblings were overseas.

7 The Applicant filed this application on 2 December 2025 following what she says were three incidents spanning the first year of the marriage, which last ended in the Respondent kicking her in the abdomen while she was holding their five-month-old infant on 30 September 2025. That incident caused her to leave the matrimonial home with [A] that night, take up residence at her mother’s home, and file a police report on 14 October 2025.

8 The three incidents are:

(a) 3 February 2025, when the Respondent allegedly shouted at the Applicant while she was seven months pregnant, over sleeping arrangements in the Respondent’s sister’s room;

(b) 6 May 2025, when the Respondent allegedly kicked aside a footrest, snatched and threw the Applicant’s mobile phone, and shouted at her during a dispute about the temperature of water used in the preparation of formula milk; and

(c) 30 September 2025, when the Respondent kicked the Applicant while holding [A].

9 The Women’s Charter (Family Violence and Other Matters) (Amendment) Act 2023 (No 21 of 2023), which came into force on 2 January 2025, brought about the most significant reform to the family violence regime since the Charter’s enactment. The amendments implemented the recommendations of the Report of the Taskforce on Family Violence released on 23 September 2021 with the aim of better protecting survivors and enhancing the rehabilitation and accountability of perpetrators.

10 The previous definition of family violence in s 64 of the Charter (now repealed) contained four limbs. These included placing a member in fear of hurt, causing hurt, wrongful confinement, and causing continual harassment with intent to cause anguish, and made no express reference to emotional or psychological abuse. The new provisions under ss 58A–60F of the amended Charter reconstitute “family violence” around three categories of abuse spanning physical, sexual, and emotional or psychological abuse. This definition aligns the Charter with the analogous provisions in the Children and Young Persons Act 1993 (2020 Rev Ed) (“CYPA”) and the Vulnerable Adults Act 2018 (2020 Rev Ed) (“VAA”), which had preceded the Charter in formally recognising emotional harm.

11 The two categories of family violence engaged in this case are physical abuse under s 58B(2) and emotional or psychological abuse under s 58B(4):

12 Two further provisions set out the scope of abuse. Section 58B(5) provides that “abuse may take the form of (a) a single instance of conduct or behaviour; or (b) a course of conduct or behaviour.” ¹ Section 58B(6) extends abuse to conduct not directly targeted at the family member, provided it is “capable of being seen, heard or otherwise perceived (directly or indirectly)” by that person. Three illustrations accompany s 58B and provide further guidance on its application:

13 Physical abuse under s 58B(2)(a)(i) is satisfied where conduct causes, or threatens to cause, personal injury or physical pain. No proof of actual injury is required. Where a respondent invokes the self-defence exception under s 58B(2)(c)(i), the force used must have been proportionate to the threat and appropriately directed to averting rather than retaliating against it: TEK v TEJ [2015] SGFC 89 (“TEK v TEJ”) at [14]–[15]; XBC v XBD [2024] SGFC 63 (“XBC v XBD”) at [11]–[12]. The self-defence exception is preventive in character. It does not sanction punitive or retaliatory force. The domestic context, including the close relationship between the parties is a relevant consideration in assessing proportionality of force employed and whether the force was appropriately used: XBC v XBD at [12].

14 Emotional or psychological abuse was unknown to the previous family violence regime as an independent basis for protection. Its express statutory recognition in s 58B(4) is a significant development that requires judicial elaboration. The practical significance of s 58B(4) has become apparent quickly. In the first year in which the new provisions were in force in 2025, emotional or psychological abuse was cited as the sole basis in approximately 32 per cent of the 1,484 PPOs issued by the Singapore courts: Theresa Tan, “‘The emotional abuse is relentless’: Dad gets personal protection order for daughter against mum”, The Straits Times (22 February 2026). These figures show that s 58B(4) is not an insignificant provision. It is evidently being actively invoked as an independent ground for protection, and in a substantial proportion of cases, it is the only ground. The need for judicial clarity on the scope of s 58B(4) and the standard of proof it demands is therefore not merely academic.

15 I observe that the social phenomenon the provision addresses is one whose boundaries are contested even in the disciplines such as psychology, social work, and sociology, which have studied it for decades. Some facts that amount to emotional or psychological abuse may at first blush appear subtle or insufficiently serious. That impression can be misleading. As then-Minister of State for Social and Family Development Ms Sun Xueling (“MOS Sun”) stated during the Second Reading of the Women’s Charter (Family Violence and Other Matters) (Amendment) Bill (Bill No 18/2023, 9 May 2023) (Singapore Parl Debates; Vol 95, Sitting No 106; [4 July 2023] (Ms Sun Xueling, Minister of State for Social and Family Development) (“Second Reading of the FV Bill”):

16 At the time of this decision, no less than six Singapore Family Court published decisions have applied or elaborated upon s 58B(4) since it came into force: see eg, XNY v XNZ [2025] SGFC 69 (“XNY v XNZ”); XWB v XWC [2025] SGFC 135 (“XWB v XWC”); XPX v XPW [2026] SGFC 30 (“XPX v XPW”); XZU v XZV [2026] SGFC 31 (“XZU v XZV”); XXW v XXX and other matters [2026] SGFC 23 (“XXW v XXX”); and YAE v YAF [2026] SGFC 35 (“YAE v YAF”). These decisions have helpfully begun to chart the contours of the provision. In these grounds, I seek to develop the analysis further, drawing in particular on a comparative survey of common law jurisdictions that have grappled with the question of what standard of assessment should be employed in their own domestic violence legislation. I do so while bearing in mind that Singapore’s legislation differs from those of other jurisdictions: see Lim Meng Suang and another v Attorney-General [2013] 3 SLR 118 (“Lim Meng Suang”) at [70]. The comparative materials are thus not cited as direct authority but as aids to reasoning about an area of law that is, across the common law world, still being developed.

17 Section 58B(4) contains two limbs that are complementary rather than cumulative. For convenience, I refer to these limbs as “Limb (a)” and “Limb (b)”, respectively. Each limb may independently establish emotional or psychological abuse, and a finding under one does not require engagement with the other:

(a) Limb (a) (“torments, intimidates, harasses or distresses a person”): This limb is directed at the character of the conduct and its impact on the person. In Public Prosecutor v Lam Leng Hung and others [2018] 1 SLR 659 (“Lam Leng Hung”), Andrew Phang JA described, at [76], the ordinary meaning of a word or phrase as the “‘proper and most known signification’ … which comes to the reader most naturally by virtue of its regular or conventional usage in the English language and in the light of the linguistic context in which that word or phrase is used.” The dictionary meaning of a word is helpful in this regard. I point out here that the terms employed in Limb (a) are drafted as verbs and not nouns. These four verbs describe distinct but possibly overlapping modes of abuse:

(i) “Torments” means “to make somebody suffer very much” (Oxford Advanced Learner’s Dictionary, Oxford University Press, online ed, entry “torment”). ² It connotes suffering of a severe kind.

(ii) “Intimidates” means “to make timid or fearful: frighten; especially: to compel or deter by or as if by threats” (Merriam-Webster Dictionary, Merriam-Webster Inc, online ed, entry “intimidate”). ³

(iii) “Harasses” means “to annoy persistently” or “to create an unpleasant or hostile situation for especially by uninvited and unwelcome verbal or physical conduct” (Merriam-Webster Dictionary, Merriam-Webster Inc, online ed, entry “harass”). ⁴

I note that a synonym of a word may be useful in the exercise of construing the ordinary meaning of a word: see Sumitomo Corp Capital Asia Pte Ltd v Salim Anthony and Other Applications [2004] 4 SLR(R) 451 at [9]. In its synonymic sense, the word “implies petty persecutions or burdensome demands that exhaust one’s nervous or mental power” (Merriam-Webster Dictionary, Merriam-Webster Inc, online ed, entry “harass”). ⁵

(iv) “Distresses” as a verb means conduct causing “pain or suffering affecting the body, a bodily part, or the mind” (Merriam-Webster Dictionary, Merriam-Webster Inc, online ed, entry “distress”). ⁶ It implies an external and usually temporary cause of great physical or mental strain and stress (Merriam-Webster Thesaurus, Merriam-Webster Inc, online ed, entry “distress”). ⁷

I would only caveat the analysis above bearing in mind the observations in Oliver Jones, Bennion on Statutory Interpretation: A Code (LexisNexis, 6th Ed, 2013) at p 1058, cited in Lam Leng Hung at [75]:

Even though the dictionary meanings of the words cannot be the end all and be all when one asks what is the proper and most known signification of the words employed in s 58B(4), they would at least provide a starting point to this inquiry.

Finally, I would also observe that in relation to the words employed, the learned DJ observed in XPX v XPW at [44(a)], Limb (a) “is framed in action terms” and focuses on “the nature of a respondent’s conduct” (emphasis added).

(b) Limb (b) (“causes or may reasonably be expected to cause mental harm to a person, including thoughts of suicide or inflicting self-harm”): This limb is directed at the consequences of conduct in terms of psychological injury. As the DJ noted at [44(b)], this limb focuses on “the consequences of the respondent’s conduct” (emphasis added). The illustrations of suicide and self-harm signal that Limb (b) is calibrated at sufficiently serious psychological harm. Mere upset, annoyance or distress, without more, will not ordinarily meet this threshold of mental harm: XPX v XPW at [44(b)].

18 The two limbs serve different but complementary purposes given the disjunctive manner in which they are drafted. Limb (a) captures conduct that causes emotional disturbance (including disturbance falling short of the requisite mental harm under Limb (b)), provided that disturbance is genuine and exceeds ordinary friction. Limb (b) is an additional way to catch cases where mental injury is severe, whether or not the character of the conduct fits neatly within the four verbs in Limb (a). A single incident may engage either or both limbs.

19 The most significant unresolved question in the interpretation of s 58B(4) which has yet to receive the attention of the High Court is the standard of assessment when determining whether conduct “torments, intimidates, harasses or distresses a person” under Limb (a). Does the court ask:

(a) whether this particular complainant was in fact tormented, intimidated, harassed, or distressed (ie, a purely subjective approach);

(b) whether the conduct would have had that effect on a hypothetical reasonable person in the complainant’s situation (ie, a purely objective approach); or

(c) something else entirely?

20 This question has practical significance because the answer determines how proof of emotional and psychological abuse operates in court in terms of what must be established, through what evidence, and to what standard.

21 I address this question in three steps. First, what the text of the provision and the parliamentary materials disclose. I am mindful that a faithful application of the text read in line with the established principles of statutory interpretation in Singapore must be the primary guidepost. Second, so far as they are relevant, I attempt also to sketch out the comparative experience of other common law jurisdictions that have come to confront squarely the manifestation of emotional or psychological abuse and its associated conduct by the name of coercive control as an aspect of family violence as understood in modern times. And third, what standard should govern under the new provision of emotional or psychological abuse under s 58B(4) in the Charter and what are the nuances, if any, to the application of this standard to each of the limbs in the provision.

22 The starting point is s 9A of the Interpretation Act 1965 (2020 Rev Ed) (“IA”), which mandates that a purposive approach be taken to the interpretation of all written law in Singapore. The provision requires that an interpretation that would promote the purpose or object underlying the written law is to be preferred over one that would not promote that purpose or object. The Court of Appeal in Tan Cheng Bock v Attorney-General [2017] 2 SLR 850 (“Tan Cheng Bock”) at [37]–[54] set out the three-step framework that a court should apply. First, ascertain the possible interpretations of the text by reference to its ordinary meaning in context; Second, ascertain the general or specific legislative purpose of the provision either from the text itself or from extrinsic materials to the extent permitted by s 9A(2)–(3) of the IA. And third, compare the possible interpretations with the legislative purpose to determine which interpretation best promotes that purpose.

23 Extrinsic materials that may be considered include explanatory statements to the Bill and parliamentary speeches recorded in the Hansard: s 9A(3)(b) and (c) of the IA. These are not used as substitutes for the statutory text, but as aids to confirm or vary an ordinary meaning arrived at from the text itself. It is this framework that I apply to the interpretation of s 58B(4) in what follows.

24 The text and parliamentary materials reveal some immediately apparent points about the specific purpose that Parliament has for the new provision.

25 First, the provision is harm-focused, not intention-focussed. MOS Sun stated in her closing speech during the Second Reading of the FV Bill that “the definition of family violence looks to the harm caused by the perpetrator, rather than the intention of the perpetrator, and this is deliberate.” The consequence is that proof of the respondent’s subjective intention to cause distress, harassment, intimidation, or torment is not required. The question is what the conduct did to the complainant or would objectively do to someone in their position.

26 Second, Limb (b) plainly partially employs an objective approach in one of its two components by its express terms. The phrase “may reasonably be expected to cause mental harm” (emphasis added in italics and bold) is plainly an objective standard. It does not require proof that mental harm was actually caused, and the adverb “reasonably” introduces the perspective of the reasonable person. I will explore the significance of the precise manner in which Limb (b) has been drafted when I set down my observations below (at [54]–[75]) on the applicable approach that should be taken for the two components in Limb (b).

27 Third, and significantly, the Explanatory Statement to the Women’s Charter (Family Violence and Other Matters) (Amendment) Bill (Bill No 18/2023, 9 May 2023) (“Explanatory Statement”) drew a deliberate distinction between conduct that distresses and conduct that is merely offensive. It states that “the definition of emotional or psychological abuse refers to conduct or behaviour that distresses an individual, and not to conduct or behaviour that is offensive to an individual”: Explanatory Statement at Clause 2(b)(iv). Clause 2(b) of the Explanatory Statement states:

28 It is evident from the Explanatory Statement that a court applying s 58B(4) does not ask whether the conduct is offensive (or unpleasant or socially unacceptable). It asks whether the conduct distressed, intimidated, harassed, or tormented the protected person, or caused or may reasonably be expected to cause that person mental harm.

29 It follows that when we look at Limb (a) which uses active verbs in the present tense (“torments, intimidates, harasses or distresses”) without the qualifier of “may reasonably be expected” in Limb (b), this admits of some uncertainty as to whether Parliament intended a subjective, objective or a hybrid standard of assessment for Limb (a). But, as I discuss below (at [48]–[53]), a purely subjective or objective reading produces practical difficulties that Parliament cannot have intended.

30 Before turning to the comparative materials, I should explain the nature and scope of the survey that follows. The foreign provisions examined below are not confined to statutory definitions of emotional or psychological abuse in terms identical or similar to s 58B(4). They span two distinct categories:

(a) First, provisions that directly define emotional or psychological abuse in family violence legislation such as Victorian state legislation and Commonwealth (federal) legislation in Australia; and

(b) Second, provisions that criminalise or regulate coercive control as a discrete form of domestic harm such as those in the United Kingdom, New South Wales, Ireland and Scotland.

31 I include both categories because they illuminate a common underlying question that s 58B(4) also raises. Coercive control is not a discrete legal concept in Singapore law, nor is it the primary lens through which s 58B(4) should be read. I elaborate on that below at [87]–[94]. What I observe at this juncture is that coercive control is one of the more extensively studied and litigated manifestations of emotional or psychological abuse in intimate relationships, and it is also a category that Parliament specifically had in mind when enacting s 58B(4) in the spousal context: Second Reading of the FV Bill. The two concepts do overlap. Coercive control is widely understood, including under s 58B(4) itself, as a form of emotional or psychological abuse. It is one defined by both its cumulative and context-dependent character and one which operates through the slow erosion of a person’s autonomy and sense of self. Several common law jurisdictions have had to grapple with coercive control precisely because it raises the same question that s 58B(4) raises, viz, when a course of conduct causes harm that is cumulative and contextual, how should a court assess whether the statutory threshold has been crossed? Because conduct amounting to coercive control would, on Singapore’s statutory provision, fall within the penumbra of s 58B(4), any local court interpreting that provision will thus inevitably have to confront questions about what coercive control means and whether particular conduct partakes of its character.

32 Next, it cannot be emphasised enough that Singapore’s s 58B(4) is, after all, a sui generis provision suited to the context of Singapore. ⁸ Its language, legislative history, and policy considerations wholly differ from those of every foreign statute surveyed below. Indeed, the High Court was clear in its observations in Lim Meng Suang ([16] supra) at [70] that:

33 I refer to these materials only as illustrations of how courts and legislatures have reasoned through that fundamental question. They are not a primary, or even secondary, guidepost to the conclusions I have reached on the proper interpretation of s 58B(4), to borrow the language employed in Lim Meng Suang. The function of the comparative survey is the more limited one of framing the interpretive issues that Singapore courts must confront. Foreign jurisprudence can illuminate the range of choices available and the difficulties of application that other systems have encountered when grappling with the same amorphous concept of emotional or psychological abuse and its other associated terms. That said, it cannot supply the answer, which must be reached independently and exclusively by reference to the statutory text, the legislative history, and the extrinsic materials permitted under s 9A of the IA, construed according to the framework in Tan Cheng Bock ([22] supra). My reasoning proceeds on that basis throughout and should not be misconstrued to be relying on the foreign legislation and authorities as either a supporting or even confirmatory basis of the conclusions I have reached.

34 The final preliminary point concerns the nature of several of the comparative materials. Some of the foreign authorities surveyed involve standards employed in the criminal context. I refer to these not because criminal standards govern here. A PPO application under the Charter is a civil proceeding and the standard of proof is the balance of probabilities: s 60A(1) of the Charter; UNQ v UNR [2020] SGHCF 21 at [24]–[28]. The experience of foreign jurisdictions is nonetheless instructive for a specific and limited purpose. Emotional and psychological abuse is both a phenomenon particular to each jurisdiction’s legal tradition and a universal problem that legislatures around the world have been called upon to address. As I will cover in this judgment, this is really the main interpretative issue confronting the local courts’ application of the new provisions under s 58B(4) of the Charter for which a range of possible interpretations may be taken. How a foreign legislature has crafted the standard of abuse (ie, by what it requires by way of impact, intent or proof) illuminates the path that Singapore courts may explore in deciding as between a subjective, objective or some other standard of assessment.

35 The UK has developed different standards for dealing with emotional or psychological abuse and its associated concept, coercive control. These reveal the full spectrum of possible approaches within that jurisdiction.

(a) The criminal standard under s 76 of the Serious Crime Act 2015 (c 9) (UK) (“UK SCA 2015”) requires that the behaviour “has a serious effect on B.” This is defined as causing B to fear that violence will be used against them on at least two occasions or causing B “serious alarm or distress which has a substantial adverse effect on B’s usual day-to-day activities” (s 76(2) UK SCA 2015). This appears from its plain wording to be a subjective assessment of the effect on B in that it requires fear or alarm or distress to have been caused. ⁹ The behaviour must also be engaged in “repeatedly or continuously.” The perpetrator’s state of mind is assessed objectively as under s 76(5), the perpetrator “ought to know” that which “a reasonable person in possession of the same information would know.” The overall requirements thus appear from their plain reading to require a mixed standard. It is subjective as to actual impact but can be objective as to the perpetrator’s state of mind, with the additional requirements of repetition and continuity that mean a single incident cannot suffice.

(b) Outside the criminal context, s 1(3) of the Domestic Abuse Act 2021 (c 17) (UK) (“UK DAA 2021”) defines “abusive” behaviour to include “psychological, emotional or other abuse”. It expressly provides that “it does not matter whether the behaviour consists of a single incident or a course of conduct”.

(c) In one of the first known reported cases in the UK Family Courts on coercive control in F v M [2021] EWFC 4 (“F v M”) (a case decided before the UK DAA 2021 came into force), Hayden J applied a holistic contextual assessment that appears to be predominantly objective in character. The learned Judge first observed (at [4]) that there was very little reported case law in the UK Family Courts considering coercive and controlling behaviour. His Honour then provided in the following terms:

Crucially, his formulation of coercive control as “a range of acts designed to render an individual subordinate and to corrode their sense of personal autonomy” (emphasis added) seems to at least consider the nature of the offender’s conduct as one part of the analysis. His observation (at [109]) that “key to assessing abuse in the context of coercive control is recognising that the significance of individual acts may only be understood properly within the context of wider behaviour” imports a contextual analysis to such conduct.

(d) The Court of Appeal (Civil Division) of England and Wales’ landmark decision in Re H-N and Others (Children) (Domestic Abuse: Finding of Fact Hearings) [2021] EWCA Civ 448 (“Re H-N”) is also instructive, for among other things, its comprehensive discussion of how the UK has grappled with the impact of family violence and its eventual encompassment of emotional or psychological abuse and coercive control as these terms have come to be known.

(i) The Court traced the evolution in the UK Family Court’s understanding of what constitutes domestic abuse. The starting point, the Court observed, was the UK’s Domestic Violence and Matrimonial Homes Act 1976, which was so narrow in its conception of “domestic violence” that it required evidence of actual bodily harm before a power of arrest could even be attached to an injunction (at [23]). Even after this statutory breakthrough, judges through the 1980s tended to treat domestic violence as a purely adult matter (at [24]). What changed the picture fundamentally was the growing recognition that it is possible to be a victim of controlling or coercive behaviour or threatening behaviour without ever sustaining a physical injury, and that specific incidents, rather than being free-standing matters, may form part of a wider pattern of abuse (at [27]).

(ii) The Court addressed Family Proceedings Rule 2010: Practice Direction 12J — Child Arrangements and Contact Orders: Domestic Abuse and Harm (“PD12J”), originally implemented in 2008, which imposes on the UK’s Family Court a mandatory step-by-step framework for cases involving allegations of domestic abuse (at [9]–[10]). The Court noted that PD12J was expanded in 2017 and now defines domestic abuse to encompass “any incident or pattern of incidents of controlling, coercive or threatening behaviour, violence or abuse,” expressly including “psychological, physical, sexual, financial, or emotional abuse” (at [26]). Crucially, PD12J defines “coercive behaviour” as acts “used to harm, punish, or frighten the victim” and “controlling behaviour” as acts “designed to make a person subordinate and/or dependent” (at [26]). Although the structure of the definition of “domestic abuse” in the Domestic Abuse Bill (or the UK DAA 2021 which has since come into force) currently differs from that in PD12J, the Court observed that the content is substantially the same. The Court found that PD12J, as expanded, remained fit for purpose and properly reflected modern understanding of domestic abuse (at [28]).

(iii) The Court then endorsed F v M, commending it as essential reading for the family judiciary (at [30]). The Court reproduced Hayden J’s central observation that coercive and controlling behaviour involves a pattern of acts whose impact must be assessed cumulatively rather than in isolation (at [29]).

(iv) The Court made clear that coercive and controlling behaviour may cause serious emotional and psychological harm regardless of whether there has been any episode of actual violence or sexual abuse. A pattern of abusive behaviour is as damaging to those within the household as any specific factual incident (at [31]).

(v) The Court was equally clear about the outer boundary of the concept. Endorsing the approach of Peter Jackson LJ in Re L (Relocation: Second Appeal) [2017] EWCA Civ 2121 at [61], the Court held that “not all directive, assertive, stubborn or selfish behaviour, will be ‘abuse’” (albeit in the specific context of proceedings concerning the welfare of a child) and that “much will turn on the intention of the perpetrator of the alleged abuse and on the harmful impact of the behaviour” (at [32]).

The standard applied by the UK Family Court as set out in Re H-N is therefore one that assesses the cumulative pattern of conduct in context, recognises psychological and emotional harm as equally serious forms of abuse as physical violence. It requires an assessment of both the nature of the conduct and its harmful impact and maintains a meaningful threshold below which ordinary relational friction does not cross into actionable abuse.

36 In Pickford & Pickford and Another (2024) 70 Fam LR 85 (“Pickford”), the Full Court of the Federal Circuit and Family Court of Australia considered comprehensively the definition of family violence under s 4AB of the Family Law Act 1975 (Cth) (“Australia FLA 1975”), which encompasses “violent, threatening or other behaviour by a person that coerces or controls a family member, or causes the family member to be fearful.” The Full Court was unanimous on two propositions of significance.

(a) First, perpetrator intent is not a necessary element of a finding of coercive or controlling conduct, though it will very often be relevant and even decisive as a factual matter (at [111] per Austin and Williams JJ; at [2] per McClelland DCJ; at [46] per Aldridge and Carew JJ).

(b) Second, and critically, the victim’s subjective perception or experience alone is insufficient. Determining whether conduct coerces or controls “requires deeper evaluation than simply a party’s experience or perception” (at [31] per McClelland DCJ; at [105], [107], [111] per Austin and Williams JJ).

37 Beyond these points of agreement, two views emerged on the structure of the definition. Austin and Williams JJ held that the two limbs demand different standards:

(a) “coerces or controls” is an objective concept — “the coercion or control must be an objective actuality”;

(b) whereas “causes the family member to be fearful” is assessed subjectively by reference to the victim’s actual reaction, without any requirement of reasonableness (at [109]–[111] and [113]–[114]).

38 The majority comprising Aldridge and Carew JJ (with McClelland DCJ agreeing) declined to limit the definition to those two types of behaviour, holding that s 4AB should not be read down by artificial limitations and that the finding of coercive or controlling behaviour is ultimately “an evaluative one, having regard to all of the circumstances and all of the evidence before the court” (at [41]–[43] and [52]). McClelland DCJ emphasised that the key to the inquiry is “an appreciation of a pattern or series of acts, the impact of which will usually be both intersecting and cumulative, rather than incident specific” (at [5]). The Full Court also endorsed the fact-finding approach articulated by Aldridge and Carew JJ, which requires the court to identify the behaviour complained of, its full context including any explanation by the alleged perpetrator and also the impact on the alleged victim. It also requires the court to explain why the behaviour does or does not amount to coercive or controlling conduct, noting expressly that “mere assertion by the alleged victim that they feel coerced or controlled is insufficient” (at [48]).

39 At the state level, the Family Violence Protection Act 2008 (Vic) (“FVPA Vic”) defines “emotional or psychological abuse” in s 7 as “behaviour by a person towards another person that torments, intimidates, harasses or is offensive to the other person.” This formulation is strikingly similar to s 58B(4)(a), save that Singapore adopts the word “distresses” in place of “is offensive to.” The Victorian provision is fundamentally conduct-focused, with no express requirement to consider the impact on the victim. In New South Wales, s 54D of the Crimes Act 1900 (NSW) creates a criminal offence of abusive behaviour towards intimate partners which requires, among other elements, that a reasonable person would consider the course of conduct would be likely, “in all the circumstances, to cause any or all of the following, whether or not the fear or impact is in fact caused”, fear of violence, or a serious adverse impact on the capacity to engage in ordinary day-to-day activities. The express words “whether or not the fear or impact is in fact caused” dispense with proof of actual subjective impact. That said, s 54D appears to be a more demanding provision than this single element suggests. It also requires the perpetrator to specifically intend the course of conduct to coerce or control the other person. This is an element absent from the text of Singapore’s s 58B(4).

40 Section 264 of Canada’s Criminal Code, RSC 1985, c C-46 (Can) creates the offence of criminal harassment. Section 264(1) reads: “No person shall, without lawful authority and knowing that another person is harassed or is reckless as to whether the other person is harassed, engage in conduct referred to in subsection (2) that causes that other person reasonably, in all the circumstances, to fear for their safety or the safety of anyone known to them.” The provision thus contains two distinct concepts: “[H]arassment”, which features as part of the knowledge (or mens rea) element, requiring the accused to know that the complainant is harassed; and second, “fear for safety”, which is the prohibited consequence the conduct must cause. It is the first concept, harassment, that is instructive by analogy to Singapore’s s 58B(4)(a).

41 In R v Sillipp (1997) 120 CCC (3d) 384 (“Sillipp”), Berger JA held that “harassment” in this sense demanded proof of something substantially more than being “vexed, disquieted or annoyed”, with the appropriate synonyms being “tormented, troubled, worried, plagued, bedeviled and badgered” (at [16]). As for the “fear for safety” consequence, the Court applied a mixed subjective-objective standard requiring both actual fear and objectively reasonable fear in all the circumstances, with the reasonableness inquiry not excluding the particular vulnerabilities of the complainant (at [18] and [27]).

42 As for family proceedings, s 1 of the British Columbia’s Family Law Act, SBC 2011, c 25 (BC, Can) (“BC FLA”) defines “psychological or emotional abuse” as part of “family violence” to include, among other things, “intimidation, harassment, coercion or threats” and expressly provides that the definition applies “with or without an intent to harm a family member.” Perpetrator intent is not necessary.

43 New Zealand’s Family Violence Act 2018 (NZ) (“NZ FVA 2018”) defines “family violence” under ss 9(1) read with s 9(2) to include psychological abuse, which is in turn defined in s 11 to include “intimidation or harassment” and threats of physical or sexual abuse, among other things. Section 9(3) further provides that violence includes a pattern of behaviour that “is coercive or controlling (because it is done against the person to coerce or control, or with the effect of coercing or controlling, the person)” (emphasis added). The provision thus creates two disjunctive bases. The first turns on purpose which looks at the nature of the conduct. There must be conduct done “because” it is to coerce or control. The second turns on the effect, regardless of the perpetrator’s intent. Section 10(2) goes further, expressly providing that a number of acts forming part of a pattern of behaviour “even if all or any of those acts, when viewed in isolation, may appear to be minor or trivial” may amount to abuse.

44 Ireland’s s 39(1) of the Domestic Violence Act 2018 (Ire) (“Irish DVA 2018”) provides that a person commits an offence of coercive control where he or she “knowingly and persistently engages in behaviour” (emphasis added) that: (a) is controlling or coercive; (b) has a serious effect on a relevant person; and (c) a reasonable person would consider likely to have a serious effect on a relevant person. Under s 39(2), a “serious effect” means the behaviour causes the person to fear that violence will be used against her, or causes serious alarm or distress that has a substantial adverse impact on their usual day-to-day activities. The Irish provision thus requires perpetrator’s intent (“knowingly”), repetition (“persistently”), actual subjective impact (“has a serious effect”), and objective confirmation of that impact by reference to a reasonable person’s point of view.

45 Scotland’s Domestic Abuse (Scotland) Act 2018 (asp 5), on the face of the statutory text, adopts an objective approach to the offence of engaging in abusive behaviour towards a partner or ex-partner. Section 1(2) provides that a further condition is whether “a reasonable person would consider the course of behaviour to be likely to cause B to suffer physical or psychological harm.” Section 4(1) then provides in unambiguous terms that “the commission of an offence under section 1(1) does not depend on the course of behaviour actually causing B to suffer harm of the sort mentioned in section 1(2).” No proof of actual impact is thus required. The provision is notable because it expressly removes the need for the complainant to establish or articulate their own psychological harm. The offence is made out by the objective likelihood of harm as assessed by the reasonable person.

46 Applying the three-step purposive framework in Tan Cheng Bock ([22] supra), I turn to the question of the appropriate standard of assessment.

(a) The first step requires the court to ascertain the possible interpretations of the text by reference to its ordinary meaning in context. The text of Limb (a) of s 58B(4)(a) (“torments, intimidates, harasses or distresses”) admits of ambiguity as to whether a subjective, objective or a hybrid standard governs.

(b) The second step requires the court to ascertain the legislative purpose (both general and specific) of the provision, from the text itself and from the extrinsic materials to the extent permitted by s 9A of the IA. As the parliamentary materials surveyed above show, the specific and not merely general purpose of s 58B(4) is to protect survivors from emotional and psychological harm assessed by reference to what was done to them rather than what the respondent intended.

(c) The third step requires the court to select the interpretation that best promotes that purpose. Of the available interpretations, I find that the Contextualised Objective Approach does so. It is objective enough to give consistent and predictable notice of what conduct is prohibited and contextualised enough to capture the relational character of emotional abuse that a purely decontextualised reasonable person standard would miss. An appropriate and workable standard for s 58B(4)(a) of the Charter is therefore what I would describe as the “Contextualised Objective Approach”, which may be described as follows: The court asks whether the conduct was, assessed in its full relational and situational context, of a character that would objectively torment, intimidate, harass, or distress a person in the complainant’s position, having regard to the specific nature and history of the relationship, the domestic setting, any known vulnerability of the complainant, the power dynamics between the parties, and the cumulative effect of prior incidents.

That said, I should qualify that this approach as framed is not a checklist. It is a description of the evaluative standard to guide the court’s assessment which must be both objective and yet sensitive to the complainant’s circumstances and relational context. The point is that the court cannot take a purely subjective or objective approach when it comes to the application of Limb (a).

47 This approach is broadly consistent with the standard adopted by the learned Magistrate in XXW v XXX and other matters ([16] supra). The Magistrate developed in his detailed analysis (at [47]–[57]) that “an objective approach should be taken in assessing whether there had been emotional and/or psychological abuse”, framed as asking “whether a reasonable person, in the position of Y, would be tormented, intimidated, harassed or distressed by X’s conduct or behaviour.” The learned Magistrate grounded this reasoning in the penal consequences of a PPO breach, that because any contravention of a PPO renders further acts of family violence a punishable offence, the standard for establishing family violence must be objective enough to give both parties and courts clear notice of what is and is not acceptable conduct: XXW v XXX at [48]–[52]. I respectfully agree with, and build upon, that reasoning. The Contextualised Objective Approach I adopt is consistent with the learned Magistrate’s formulation but makes explicit what his approach implies, which is that the reasonable person must be placed in the complainant’s actual relational position, with full regard to the history and dynamics of the specific relationship and not assessed as a generic bystander divorced from context. This appears to have been what the Magistrate intended when he described the approach as the “objective approach”, as he later went on to hold (at [57]) that “one must also have regard to the circumstances of the case (including Y’s particular characteristics), as well as the relevant context in which the alleged conduct or behaviour had taken place”. This qualification of the objective approach matters because the same conduct, such as turning off a fan without explanation or the forceful throwing of a bottle at a wall may carry an entirely different complexion depending on what has preceded it between these particular parties. The Contextualised Objective Approach preserves the objectivity required by the penal consequences argument while ensuring that context, which is central to how emotional and psychological harm actually operates, is not stripped away.

48 I arrived at this formulation by reasoning through the difficulties presented by the two pure alternative approaches under a purely subjective or objective approach.

49 A purely subjective approach which asks only whether this particular complainant was in fact distressed, intimidated, harassed, or tormented, is problematic for two related reasons. The first, as the Magistrate identified in XXW v XXX at [48]–[52], is that it provides no objective notice of what conduct is prohibited. Whether a respondent’s act crosses the threshold would turn entirely on the internal psychological state of the particular complainant. This is not something the Respondent can claim to really know in every case, which may vary from person to person, and which gives courts no consistent basis for decision. That sits uneasily with the penal consequences that attach to a breach of a PPO, which demand a standard capable of giving clear and consistent notice to both parties of what is and is not acceptable conduct. I respectfully agree with the Magistrate’s reasoning on this point.

50 The second difficulty is what I would describe as practical, if I may put it that way. Limb (a) employs the four verbs “torments, intimidates, harasses or distresses” which, on their face, describe a wide spectrum of possible emotional states. A purely subjective standard would open the provision to claims by anyone who has subjectively experienced distress, intimidation, harassment, or torment as a result of a respondent’s conduct, however unreasonable or idiosyncratic that reaction may be. This concern is sharpened when Limb (a) is compared with Limb (b). As I explain below (at [67]), Limb (b) contains its own internal safeguard. It is anchored in mental harm, ie, harm serious enough to cause thoughts of suicide or self-infliction of injury. The illustrations of suicide and self-harm signal that Limb (b) operates at the more severe end of psychological injury, and as I discuss in the context of Limb (b) below at [65(a)], the causation inquiry under that component of the limb is one established through objective evidence of mental harm rather than through subjective self-reporting alone. Limb (a) contains no equivalent explicit severity marker. If its terms are interpreted purely subjectively, the provision risks being invoked in respect of conduct that any robust relationship must be expected to weather. These may include the unkind word, the argument that goes too far, the silence that persists too long. A contextualised objective standard prevents this by requiring the court to assess whether a reasonable person in the complainant’s position would have been genuinely tormented, intimidated, harassed, or distressed, not merely whether this particular complainant says she was.

51 A third difficulty with the purely subjective approach is more fundamental and goes to the logical structure of the provision itself. On a proper reading of s 58B(4), Limb (a) and Limb (b) are directed at different inquiries. As the learned DJ observed in XPX v XPW ([16] supra) at [44], Limb (a) “focuses on the nature of a respondent’s conduct” and “is framed in action terms: ‘torments, intimidates, harasses or distresses’”, while Limb (b) “focuses on the consequences of the respondent’s conduct.” Assuming that this distinction is one which is correct, this distinction has a logical implication. If Limb (a) is, at its core, concerned with the character of the respondent’s conduct, asking, in effect, what the conduct did or was capable of doing then the frame of inquiry is, by its nature, conduct-facing rather than complainant-facing. The question is not primarily what the complainant felt, but what the conduct was. A purely subjective standard, which asks only whether this particular complainant was in fact distressed, would invert this assessment entirely by reorienting the inquiry away from the conduct and toward the complainant’s internal psychological state. This would effectively collapse the distinction between Limb (a) and Limb (b) and render it otiose. In my judgment, the statutory canon of interpretation that “that the thing may rather have effect than perish” (ie, the principle of effectiveness as encapsulated by the Latin maxim ut res magis valeat quam pereat) is applicable here and it requires that each limb be given an interpretation that is distinctly significant. A purely subjective standard for Limb (a) would deprive it of that distinctiveness by making it less functionally distinguishable and would effectively become a weakened version of the consequence inquiry under Limb (b). I do not think this can be right, but of course, this argument is only sound if the distinction in XPX v XPW at [44] is correct, which in my respectful view, it is based on the manner in which Parliament has drafted s 58B(4) and for the reasons I have sketched out above.

52 Next, the alternative purely objective approach of asking only whether a generic reasonable person would have been affected, without regard to the specific relational context, faces the opposite difficulty. It strips away precisely the context that gives emotional or psychological abuse its character. Whether a particular act is intimidating depends heavily on what preceded it in this relationship, on the power dynamics between these parties, and on the complainant’s specific vulnerabilities. Conduct that appears unremarkable to a stranger may be profoundly intimidating when viewed against a history of aggression. A purely decontextualised reasonable person standard would fail to capture the most characteristic forms of emotional harm, which derive their coercive force precisely from parties’ existing relational history.

53 In my judgment, the Contextualised Objective Approach, and this is one which many of my sister and brother judges have come to apply, best resolves both difficulties. It preserves the objectivity required by the penal notice argument identified in XXW v XXX ([16] supra) and prevents the floodgates concern by anchoring the assessment in what a reasonable person in the complainant’s position would have experienced. It simultaneously contextualises that reasonable person by placing them in the complainant’s actual relational circumstances, ensuring that the ordinary-seeming act is assessed against the relational backdrop that gives it meaning. The personal sensitivity that no reasonable person would share is excluded from this inquiry. That said, the harm that a reasonable person in the complainant’s position would have actually suffered must also be considered.

54 A fundamental interpretive question confronts the court in reading Limb (b). The provision states that emotional or psychological abuse means conduct that “causes or may reasonably be expected to cause mental harm to a person, including thoughts of suicide or inflicting self-harm.” On its face, the text raises two potentially distinct triggers for s 58B(4) to apply as suggested by the use of the word “or”: (a) conduct that causes mental harm; or (b) conduct that may reasonably be expected to cause it. In interpreting s 58B(4), this Court must thus address which of the following interpretations should prevail:

(a) Limb (b) is to be read as two disjunctive components governed by different standards of assessment; or

(b) Limb (b) is to be read such that a single uniform standard (ie, the Contextualised Objective Approach) applies across the entirety of Limb (b) without distinction.

55 The difficulty of this question is underscored by the experience of Australia, where the Full Court of the Federal Circuit and Family Court of Australia (Division 1) ¹⁰ as a superior court for family law matters in Australia was required to confront a similar interpretative problem. Even that court found itself divided. That the Full Court in Pickford ([36] supra) was called upon to determine whether a provision structured in disjunctive terms demanded a single uniform standard or two different ones, and that it produced a plurality of opinions, is itself indicative of the difficulty. For that reason, the Full Court's treatment of this question illuminates the terrain that our courts have to navigate under s 58B(4).

56 In Pickford ([36] supra), the Full Court of the Federal Circuit and Family Court of Australia confronted the interpretation of Australia’s definition of “family violence” under s 4AB of the Australia FLA 1975 that contained one limb directed at the nature of the conduct (ie, behaviour that “coerces or controls”) and another directed at its consequence upon the victim (ie, behaviour that “causes the family member to be fearful”):

57 In addressing the proper interpretation of this provision, the following passage from Austin and Williams JJ’s judgment at [109] in referring to the definition as set out at [55] above, addressed the interpretive choice they confronted:

58 Austin and Williams JJ held that the “causes fearfulness” limb, focusing as it does on the victim’s reaction rather than the nature of the behaviour, called for a different standard from the objective one governing the coercion and control limb. In Their Honours’ view, this standard should be one described to be “subjective”. In contrast, the majority in Pickford comprising Aldridge and Carew JJ, with McClelland DCJ agreeing (at [3]), took a broader view of the definition above. Their Honours declined to limit “family violence” to only two types of behaviour (ie, that which coerces or controls and that which causes fear) and instead opted to treat the overall inquiry as an evaluative one having regard to all the circumstances (at [41]–[43] and [52]). This is a broader holistic approach that the majority preferred.

59 The divergence of views in Pickford serves to illustrate that two principal approaches are available to a court grappling with a provision structured as s 58B(4) is. The first is a uniform approach applying the Contextualised Objective Approach across both components of Limb (b) without distinction. I observe that this would be more aligned with the majority’s view in Pickford. The second is a differentiated approach recognising that the first component, which asks what actually happened to this complainant, calls for a different standard from the second which instead asks what a reasonable person in the complainant’s position would be expected to experience in future. For the avoidance of doubt, I have illustrated these two possible interpretative approaches through the reasoning in Pickford not because the reasoning in that decision can apply here, but because the Full Court there had occasion to confront the same interpretative issue, indeed, they do illuminate the interpretive choice that local courts must make.

60 In my view, based on the text of s 58B(4)(b), Limb (b) should be read as comprising two disjunctive components that call for the application of different standards. Parliament drafted Limb (b) in two disjunctive ways when it employed “causes” for actual harm that has occurred, and “may reasonably be expected” for harm that has not yet occurred but is anticipated. In other words, the use of the past-present tense for “causes” suggests a consequence that has already materialised (or at the maximum, continuing). The use of what I would describe as future-facing language for “reasonably expected” points to an anticipated one. In my judgment, different approaches must be taken for either component because to apply the same standard to both would be to read “causes” as if it meant “would reasonably be expected to cause”. This would render the distinction Parliament expressly drew otiose, contrary to the principle of effectiveness (known also as ut res magis valeat quam pereat principle). Parliament shuns tautology and does not legislate in vain such that the court should therefore endeavour to give significance to every word in an enactment: Tan Cheng Bock ([22] supra) at [38], citing JD Ltd v Comptroller of Income Tax [2006] 1 SLR(R) 484 at [43].

61 I observe that the disjunctive interpretation of Limb (b) aligns more with (but is not independently supported or confirmed by) the minority view of Austin and Williams JJ in Pickford. Should the view that a single uniform holistic inquiry be applied across s 58B(4) be adopted? I raise a contrast with the majority’s view in Pickford, not in support of my view above (this is impermissible insofar as I am interpreting a Singapore statute), but to raise some questions which a future court may have to resolve more directly with the benefit of fuller submissions. The majority comprising Aldridge and Carew JJ, with McClelland DCJ agreeing (at [3]), declined to accept the minority’s bifurcation of the Australian definition into only two types of behaviour, viz, an objective “coerces or controls” limb and a subjective “causes fearfulness” limb. Their view was that s 4AB of the Australia FLA 1975 should not be read down by artificial limitations (at [43]), that the overall inquiry is an evaluative one having regard to all the circumstances (at [52]), and that determining whether conduct coerces or controls “requires deeper evaluation than simply a party’s experience or perception” (at [31] per McClelland DCJ). In my view, the true question for a local court, is really not one of selecting as between the minority’s or the majority’s view in Pickford, but whether the subjective approach (which the minority there adopted under the Australian provision for the “causes … fearful” limb) is appropriate for Singapore’s analogous “causes mental harm” component in s 58B(4). This is an inquiry that our local courts have to answer with reference to our own statutory text and other permissible interpretative aids.

62 Apart from the textual reasons I laid out above at [60], I would proffer an additional reason why a subjective approach is feasible and even preferred when construing Limb (b) of s 58B(4). Singapore’s Parliament has expressly built a seriousness threshold into the provision itself, in the form of the illustrative examples of thoughts of suicide and self-harm. A complainant who cannot establish harm of that serious character will not satisfy Limb (b) regardless of how their account is framed. The subjective inquiry is therefore appropriate in Singapore precisely because the demanding seriousness threshold does the work that any objective standard would otherwise need to do. I would only observe in passing that this same safeguard does not appear to have been built into s 4AB of Australia’s FLA 1975 which plainly only requires that the behaviour “causes the family member to be fearful” (see above at [56]). Singapore’s statutory framework thus makes the subjective approach not only defensible, but also the more natural reading of the provision as Parliament enacted it. I elaborate more on the seriousness threshold under Limb (b) in greater detail at [67] below.

63 Assuming that a disjunctive reading should be given to Limb (b), I turn then to address the different standards of assessment that apply to each of the two components of Limb (b).

64 In my respectful view, the proper standard of assessment for the first component (“causes … mental harm”) is to ask: has mental harm of the requisite kind actually occurred, and was it caused by the respondent’s conduct? This approach has been described to be “subjective” by Austin and Williams JJ in Pickford ([36] supra) at [108]–[109], but I would not term the approach to be a subjective standard of assessment in the strict sense. It may be more appropriately described as a factual inquiry directed at what actually happened to the complainant, which is an objective finding of fact, at a higher level of abstraction. The element of subjectivity lies in the nature of the harm itself (this being “fearful” in the Australian context, but “mental harm” in the Singapore context) which can come from the complainant’s own account.

65 In other words, the court must determine, on a balance of probabilities on all the evidence, whether the requisite harm occurred and whether the respondent’s conduct caused it. The “causes” component is ultimately backward-looking. Two sub-questions arise under this inquiry:

(a) The first is causation. Did the respondent’s conduct cause the harm? This is the familiar inquiry of factual causation which courts undertake across many fields of law. The court assesses the evidence as a whole taking into consideration the complainant’s account, the chronology of events, the nature of the conduct, and any corroborating material, and determines whether the causal link is established. This is an objective assessment in the sense that the court is not bound simply to accept the complainant’s assertion, but it is a factual assessment, not a normative one. The court is finding what happened, not what a hypothetical reasonable person would have experienced.

(b) The second sub-question is the nature and seriousness of the harm. The mental harm required under Limb (b) is not ordinary unhappiness, frustration, or sadness. Applying the well-established statutory canon of interpretation that a word is known by its associates (ie, the Latin maxim noscitur a sociis), “mental harm” must be of a character approaching the gravity of the illustrative examples Parliament chose, which would be thoughts of suicide and thoughts of inflicting self-harm. These examples are telling in a way that deserves attention. Parliament deliberately adopted the word “thoughts.” On this footing, the application of an objective standard may not be fully compatible in assessing a person’s thoughts. A person has had such thoughts or she has not. The inquiry is whether these particular mental consequences (or harm of analogous severity) actually materialised. In my judgment, it is not whether a reasonable person in the complainant’s position would have had such thoughts.

66 This is important for how the first component operates in practice, particularly in PPO proceedings where parties are frequently unrepresented and without access to psychiatric services. A complainant who has in fact experienced thoughts of suicide or self-harm as a consequence of the respondent’s conduct will in most cases be able to establish this only through their own testimony.

67 The safeguard against a complainant too readily satisfying the first component lies not in incorporating an objective standard of assessment, but in the seriousness threshold for the kind of mental harm required. Parliament required “mental harm” of a character illustrated by thoughts of suicide and self-harm. These are harms that are plainly grave and serious. Ordinary distress or transient upset does not meet this threshold. A complainant who cannot establish harm of that serious character will not satisfy the first component, regardless of how their account is framed. This is a meaningful and demanding standard. In this respect, the learned DJ’s analysis in XPX v XPW ([16] supra) at [44(b)] that Limb (b) focuses on “the consequences of the respondent’s conduct” and is “calibrated at the more serious end of the spectrum of psychological harm” is directly applicable. The nature of the consequences matters. What does not arise as a separate operative question under the first component is how the harm should have appeared through the eyes of a reasonable observer. The first component asks what actually happened, not what would have been expected to happen.

68 By way of passing comparison, and bearing in mind that Singapore’s statute differs from foreign provisions, the seriousness threshold built into Singapore’s Limb (b) bears resemblance to Ireland’s approach under s 39(2) of the Irish DVA 2018, which similarly requires that conduct cause “serious alarm or distress” with “a substantial adverse impact on [the complainant’s] usual day-to-day activities.” Both provisions, in their different ways, set a floor below which less serious harms do not suffice.

69 A point that has to be decided in a future case is what then is the threshold at which “mental” harm under Limb (b) should be pegged? I provide some observations on this in passing, without deciding the point, for the consideration of future courts. The structure of the CYPA draws a distinction between “emotional or psychological abuse” and “emotional harm”. The latter is the trigger for a finding that a child or young person is in need of care and protection. Section 5(1)(g) of the CYPA provides that a child or young person is in need of care and protection where he or she “suffers or is likely to suffer from emotional harm because the child or young person has been or is subject to emotional or psychological abuse by his or her parent or guardian.” “Emotional harm” is separately defined in s 2(1) to mean “any serious impairment to the growth, development, or behavioural, cognitive or affective functioning” of the child or young person and includes delayed mental and physical development, a child assessed by a relevant person under the Act (eg, Director-General, protector or approved welfare officer) to be a danger to himself or to other persons, to be severely withdrawn, anxious or depressed, and a child or young person diagnosed by a medical practitioner with a mental health condition such as post-traumatic stress disorder, anxiety, depression or psychosomatic disorder. The CYPA thus treats emotional harm not as co-extensive with emotional or psychological abuse, but as a further consequence that must be established before state intervention is warranted. This conduct and consequence distinction parallels, to some degree, the structure of s 58B(4) of the Charter where Limb (a) focuses on the nature of the conduct and Limb (b) focuses on the consequence (ie, mental harm) that results or is expected to result. It is open to a future court to consider whether the “mental harm” in Limb (b) of s 58B(4), read alongside the CYPA’s definition of emotional harm, should take its character from a similar level of seriousness. These appearing to be harm that causes significant impairment to the person’s functioning, approaching the kind of harm described in the CYPA’s non-exhaustive list which extends not merely to the child’s own person but possibly extends to danger to others. As this should be dealt with in a future court where this squarely arises, I say no more on this.

70 One objection to adopting the above factual approach to the first component of Limb (b) deserves acknowledgment. If the first component of Limb (b) is assessed on a purely factual basis by asking whether this complainant actually suffered thoughts of suicide or self-harm, does this not open the door to unmeritorious claims by self-reporting? The concern has some force. A complainant who wishes to satisfy this component need only assert that she had such thoughts, and in many cases their testimony will be the only direct evidence available.

71 But I do not think this concern is sufficient to displace the approach I have described, for two reasons. The first is that the concern is ultimately one of proof, not of standard. The question whether a complainant has in fact suffered the requisite harm is determined on the balance of probabilities on all the evidence before the court. A bare assertion, unsupported by any corroborating circumstance such as no medical or psychiatric records, no contemporaneous communications, no account of the specific thoughts experienced, no behavioural evidence consistent with serious psychological harm will carry little weight. Courts assess credibility and consistency as a matter of course. The mere possibility of self-reporting does not justify imposing a standard that denies protection to those who genuinely satisfy it.

72 The second reason is more fundamental and speaks to the mischief that Parliament sought to address in the new provisions. The alternative of requiring a complainant who has actually been driven to thoughts of suicide or self-harm to demonstrate that an ordinary person in their position would have experienced the same would deny protection to the very class of harmed persons Parliament sought to protect. From its plain terms, Parliament crafted Limb (b) to protect those who have been caused such serious mental harm. A complainant whose psychological constitution made them more vulnerable be it from prior trauma, a pre-existing condition, or maybe even the cumulative weight of an abusive relationship should not find the court’s door closed to them simply because a person of ordinary fortitude might have fared better. The respondent must take the complainant as he finds her where serious mental harm has actually resulted. To say otherwise would not protect against false claims. Instead, it would exclude genuine claims. To my mind, Parliament did intend for the respondent to take the complainant as they found them. To borrow a convenient analogy only for illustration, this is similar to the eggshell skull principle applied in other areas of law such as tort law. The court’s concern is not whether an ordinary person would have suffered the same serious harm, but whether this particular complainant actually did.

73 I turn to the second component of Limb (b).

74 The second component is forward-looking. It asks whether the conduct may reasonably be expected to cause mental harm to the complainant? Proof that harm actually resulted is not required. The word “reasonably” expressly imports an objective standard and Parliament’s choice of the conditional “may” also pitches the inquiry at one of anticipated probability, not as an actuality suffered by the complainant before the court. Again, bearing in mind the differences in statutory language and legal context, the second component’s structure bears a passing resemblance to Scotland’s approach under s 1(2) of the Domestic Abuse (Scotland) Act 2018, which asks whether “a reasonable person would consider the course of behaviour to be likely to cause B to suffer physical or psychological harm”, with s 4(1) expressly providing that the offence “does not depend on the course of behaviour actually causing B to suffer harm.” Both provisions appear to employ a forward-looking objective standard that dispenses with proof of actual harm.

75 For this second component, the Contextualised Objective Approach should apply. When the question is whether harm would be expected to result in the future, that question cannot be answered without some form of comparator. A court cannot determine whether mental harm of the kind under Limb (b) is reasonably to be expected without asking itself: to be expected by whom and in what circumstances? The answer should be the reasonable person placed in the complainant’s actual circumstances, with their known vulnerabilities, the history of the relationship, the domestic context, and the cumulative effect of the respondent’s conduct. Without this contextualised reasonable person, the inquiry would either collapse into a purely subjective assessment (ie, would this particular complainant say she fears future harm?) or a purely decontextualised one (would a generic reasonable person or “man on the street” suffer mental harm?). Both alternatives should not apply for the very same reasons why both approaches would not work for Limb (a) as stated above at [49]–[52]. More specifically, a purely objective approach of asking only whether a generic reasonable person would have been affected, without regard to the specific relational context, strips away the context that gives emotional or psychological abuse its character. Whether a particular act is intimidating depends heavily on what preceded it in this relationship, on the power dynamics between these parties, and on the complainant’s specific vulnerabilities. Conduct that appears unremarkable to a stranger may be profoundly intimidating when viewed against a history of aggression.

76 Having dealt at length above with the question of which limb applies and the standard governing each, I turn now to address how this Contextualised Objective Approach should apply to s 58B(4) as can be gleaned from the existing case law and Parliamentary materials. Three features of this Contextualised Objective Approach merit emphasis.

77 First, the perpetrator’s intent is not necessary. This is confirmed by express statement made at the Second Reading of the FV Bill and follows from the structure of the provisions which are focussed on the harm. This is also the conclusion reached in XXW v XXX ([16] supra) at [71]–[75], where the learned Magistrate held that the amended definition of family violence no longer requires proof of intention or knowledge on the part of the respondent, while acknowledging that such intention or knowledge, where it exists, remains relevant to evaluating the nature, purpose and context of the alleged conduct, and to the necessity assessment. The more appropriate question is what the conduct did or would objectively do, not what the respondent intended.

78 For comparative reference, Singapore is broadly aligned on this point with the jurisdictions surveyed at [30]–[45] above. For instance, in the UK, both under the Family Court’s approach there as applied in F v M ([35(c)] supra) and under s 1(3) of the UK DAA 2021, which contains no intent requirement at all (see above at [35(b)]). This is also supported by the legislation in Victoria (FVPA Vic), Canada (BC FLA), New Zealand (NZ FVA 2018) and Scotland (Domestic Abuse (Scotland) Act 2018). Finally, in Australia, it was the unanimous position in Pickford ([36] supra) that perpetrator’s intent is not a necessary ingredient of a finding of coercive or controlling behaviour but may very often be relevant and even decisive as a factual matter (see above at [36(a)]). It remains an open question whether Singapore courts would hold the same as the Court did in Pickford that such intent may even be decisive when it squarely confronts the issue of coercive control.

79 Second, actual demonstrated impact is probative where it is present, but it is not necessary for Limb (a) and the second component of Limb (b). Where a complainant proves through consistent and credible evidence including medical or psychiatric evidence of psychological harm that she was in fact tormented, intimidated, harassed, or distressed by the conduct, such impact is probative in satisfying Limb (a) and the second component of Limb (b): see XNY v XNZ ([16] supra) at [27]–[29] on the burden of proof for s 58B(4) allegations. This is provided of course that the impact rises above the ordinary friction inherent in any close relationship. These propositions flow from the application of the Contextualised Objective Approach to Limb (a) and the second component of Limb (b), as the approach has an objective complexion in that the complainant’s actual experience would inform the inquiry but does not alone determine the inquiry. The ultimate question is whether a reasonable person in their position would have been similarly affected: see XXW v XXX at [54]–[57]. In contrast, where the evidence shows that the alleged victim displayed no signs of genuine fear or distress for instance, by behaving in a spirited, defiant, or otherwise unaffected manner this has been taken to count against a finding: see eg, XWB v XWC ([16] supra) at [26] and [40]; XPX v XPW ([16] supra) at [46]; XZU v XZV ([16] supra) at [44].

80 I would observe that in the context of the second component of Limb (b) specifically, actual impact is likely to be of particular practical importance, even though proof of actual harm is not formally required. The second component asks whether the conduct may reasonably be expected to cause mental harm of the serious kind described in Limb (b), which is set at a high threshold approaching thoughts of suicide or infliction of self-harm. To satisfy a court that such grave consequences are reasonably to be expected from the respondent’s conduct, evidence that the complainant has actually suffered a not insignificant level of harm such as the corresponding types of harm under Limb (a) of being tormented, intimidated, harassed, or distressed, or more serious types of harm approaching or exceeding those listed under Limb (b), would conceivably carry greater weight in the court’s assessment. Conversely, the complete absence of any impact may make it correspondingly more difficult to establish that harm of that character is objectively to be expected to be caused.

81 Third, context is everything. The significance of individual acts may only be properly understood within the broader relational context. This follows naturally from the Contextualised Objective Approach. That context matters to the assessment under s 58B(4) is evident in the Family Courts’ decisions thus far. In XXW v XXX ([16] supra), the learned Magistrate observed expressly at [57] that in applying the objective standard, “one must also have regard to the circumstances of the case (including Y’s particular characteristics), as well as the relevant context in which the alleged conduct or behaviour had taken place.” In XNY v XNZ, the same learned magistrate found distress established by examining the specific relational context, viz, the close mentor-mentee bond between the applicant and his former employer, and the applicant’s religious background, and concluded at [28]–[29] that the false accusations of black magic were distressing precisely because of those particular circumstances. In XWB v XWC [2025] SGFC 135 at [30], the learned DJ assessed the father’s remarks against “the broader context of the ongoing argument between the children” in which they were made. Further, in XPX v XPW, the learned DJ asked at [45] whether the conduct had the character alleged “when viewed objectively and in context.” The Contextualised Objective Approach simply articulates what our courts have consistently been doing in practice.

82 The approach under the Charter is well aligned with those in the jurisdictions surveyed. To borrow from Hayden J’s words in F v M ([35(c)] supra) at [109], where His Honour stated within the context of assessing coercive and controlling behaviour in family proceedings, “the significance of individual acts may only be understood properly within the context of wider behaviour” (see also Re H-N ([35(d)] supra) at [29]).

83 The Australian Court in Pickford ([36] supra) likewise identified (at [5], per McClelland DCJ) the proper approach as assessing “a pattern or series of acts, the impact of which will usually be both intersecting and cumulative, rather than incident specific.” While a single incident can constitute emotional or psychological abuse under s 58B(5), where there are multiple incidents, each must be assessed both individually and as part of the cumulation of acts. By way of further comparison, the New Zealand legislature even expressly provides in s 10(2) of the NZ FVA 2018 that even “minor or trivial” acts forming part of a pattern may constitute abuse. This is an intuitive position and one which is consistent with the approach under s 58B(4) of the Charter.

84 Not every act causing unhappiness, frustration, or offence in a relationship constitutes emotional or psychological abuse. The Explanatory Statement makes this amply clear. The Family Courts in Singapore have also been consistent in recognising this. In XWB v XWC ([16] supra), the learned DJ observed at [11] that “the statute uses ‘emotional abuse’ rather than ‘emotional distress’, suggesting the victim must suffer some level of emotional harm beyond mere unhappiness” (emphasis added in bold and italics). In another case in XZU v XZV, the learned DJ held at [39] that the harm “has to go beyond ordinary feelings of frustration, indignation, annoyance and unhappiness which is inherent in everyday life.” This view finds further support in XXW v XXX at [87], where the learned Magistrate found that a father who had walked away from an incident with a “bruised ego” and was “upset or even aggrieved” had not crossed the threshold. The conduct instead amounted to ordinary friction rather than torment, intimidation, harassment, or distress in the statutory sense.

85 I respectfully agree with all these statements of my brother and sister judges insofar as they suggest that not every act causing unhappiness, frustration, or even offence in a relationship constitutes emotional or psychological abuse, but that something more is required. I set out shortly below at [87]–[94] a discussion on the threshold the court should employ to determine what qualifies as emotional or psychological abuse and more specifically, whether it has to amount to coercive control before it can amount to such abuse.

86 I observe only by way of comparison that the Court of Appeal of England and Wales in Re H-N at [32] has very much expressed the same point in the context of their own family law in stating that “not all directive, assertive, stubborn or selfish behaviour will be ‘abuse’.” In Sillipp ([41] supra), the Alberta Court of Appeal required that the complainant be “tormented, troubled, worried, plagued, bedeviled and badgered”, not merely “vexed or annoyed”, albeit in the context of their criminal harassment legislation.

87 There is a question of considerable importance that has yet to receive sustained judicial attention: does the link between s 58B(4) and coercive control mean that a complainant must establish conduct of a coercive or controlling nature before the provision is engaged? Or is it sufficient to establish that the conduct tormented, intimidated, harassed, or distressed the complainant or that it caused or may reasonably be expected to cause mental harm, without more? The answer to this question has real implications for how the statute is applied. If coercive control is the threshold, then a victim of sustained torment, intimidation, harassment, or distress that does not fit the paradigm of a controlling relationship may find herself without recourse. If the statutory terms (ie, “torments, intimidates, harasses or distresses” or “causes or may reasonably be expected to cause mental harm to a person, including thoughts of suicide or inflicting self-harm”) are sufficient on their own, then the provision operates as Parliament enacted it, broadly, and without further limitation by a concept that Parliament chose not to include into the text. Judicial guidance on this precise question is, as yet, scarce.

88 Two distinct strands of case law have emerged in the short period since the new provisions came into force.

(a) The first strand as reflected in XPX v XPW ([16] supra), focuses squarely on whether the conduct in question satisfies the terms of each limb, ie, whether it tormented, intimidated, harassed, or distressed the complainant within Limb (a), or whether it caused or may reasonably be expected to cause mental harm within Limb (b). On this approach, the statutory descriptors are really the test themselves. The question is whether the conduct meets them, assessed in its full relational and situational context. The assessment of seriousness operates at the level of each limb itself and, separately, at the necessity stage.

(b) The second strand, as suggested by passages in XWB v XWC at [10] and [34] and YAE v YAF ([16] supra) at [9], seems to treat the “egregious forms of controlling behaviour” referred to in the Second Reading of the FV Bill as the operative threshold. This may not have been the intention of the learned Judges in those cases, but these cases may arguably be relied on for the proposition that conduct which does not partake of a controlling or coercive character may not cross the statutory threshold even where it intimidates, harasses, or distresses the complainant.

I must preface the discussion below that these cases were correctly decided on their facts. In relation to YAE v YAF, the PPO application in that case was dismissed on facts where the respondent had, among other things, attempted to add the applicant’s friends on social media and made reputational complaints to mutual acquaintances. On those facts, the result is entirely defensible as the conduct in question was insufficiently serious to constitute emotional or psychological abuse under any formulation, and the dismissal of the application is well-supported by the evidence as described in the grounds. In relation to XWB v XWC, the applications were likewise dismissed after a careful analysis of multiple incidents in the context presented, and the result in that case is similarly defensible on the evidence.

89 To the extent that the decisions in the second strand of case law lay down the broader proposition that s 58B(4) is confined to “egregious forms of controlling behaviour” such that all emotional or psychological abuse must partake of the character of coercive control, it is my respectful view that this is not what s 58B(4) requires. I provide four reasons.

90 First, a coercive control standard sits in tension with s 58B(5), which expressly provides that a single instance of conduct or behaviour may constitute abuse. Coercive control is by nature a phenomenon involving a cumulative series of conduct, requiring, if I may illustrate this through Hayden J’s analysis in F v M ([35(c)] supra) without deciding that this is necessarily the law in Singapore, detailed particularisation of an extensive course of conduct and a holistic assessment of the cumulative effect of such conduct. In other words, this entails a forensic exercise across multiple incidents over time. A provision anchored in coercive control sits uneasily alongside one that can expressly be triggered by a single act.

91 Second, Parliament deliberately removed any requirement to prove the perpetrator’s intent. The “wilfully or knowingly” fault elements in the old s 64 are absent from s 58B(4). A coercive control standard reintroduces intent considerations by the back door, since coercive control as treated in the comparative jurisdictions surveyed above characteristically requires conduct aimed at establishing dominance over the complainant. This is precisely what Parliament chose not to require. As held in XXW v XXX ([16] supra) at [71]–[75], intent or knowledge is no longer a prerequisite, though it remains relevant context.

92 Third, coercive control is better understood as a circumstance that contextualises findings under Limb (a) or Limb (b) rather than as the test for those findings. Where a pattern of coercion and control is established, it provides strong evidence that specific acts within that pattern tormented, intimidated, harassed, or distressed the complainant, and explains why conduct unremarkable in isolation carries different weight when assessed against that relational backdrop. I observe without deciding the point that coercive control may also be directly relevant to necessity. Where domination is established, the risk of future family violence is typically higher, strengthening the case for a PPO. These are the evidentiary and contextual functions of an assessment of coercive control, but they do not require coercive control to serve as a legal threshold.

93 Fourth, the parliamentary materials, read in their full context, reveal that the provision was designed to be broad. In this connection, I raise the following observations based on the Second Reading of the FV Bill.

(a) MOS Sun framed emotional and psychological abuse in the opening of her speech in these terms:

These are not statutory provisions and they should not be approached as one would approach the interpretation of the same with the kind of precision expected from Parliament’s drafted words. But two observations flow from MOS Sun’s passage. The phrase “such as” is in the language of illustration, not definition. Parliament was giving examples of emotional and psychological abuse, not exhaustively defining its outer boundary. More significantly, the two examples chosen, ie, exerting control over another’s behaviour, and victim blaming, are analytically distinct. Victim blaming does not involve control or dominance in any technical sense. It is a form of emotional manipulation or humiliation that may occur in a single interaction and without any pattern of domination. By placing victim blaming alongside controlling behaviour as co-equal examples, Parliament confirmed at the very outset that the category of emotional and psychological abuse is broader than coercive control and encompasses the full range of conduct that inflicts genuine psychological harm on a person in a family relationship.

(b) The passage that is most directly relied upon in support of a narrower reading to import a requirement of coercive control does not bear the weight placed upon it:

From this excerpt, Parliament stated that controlling behaviour “falls within” the definition of emotional or psychological abuse and not that it constitutes the definition. The provision is not defined as controlling behaviour. Rather, controlling behaviour is identified as one category of conduct that the provision captures. The words “would be considered emotional or psychological abuse” confirm that Parliament is classifying the controlling behaviour examples under a broader statutory head, not stating that those examples are coextensive with it. To treat this passage as a definition which circumscribes the provision to “egregious forms of controlling behaviour” would be to read Parliament’s illustration as Parliament’s enactment. That is a step the purposive approach to statutory interpretation does not permit as extraneous materials can only confirm but not alter the ordinary meaning of the provisions as purposively ascertained: Attorney-General v Ting Choon Meng and another appeal [2017] 1 SLR 573 at [59(c)].

(c) Third, in her closing speech, MOS Sun, responding to Member of Parliament for Chua Chu Kang Mr Zhulkarnain Abdul Rahim’s question about how the updated definition should be interpreted, stated:

This formulation focuses on harm to the survivor as the touchstone of the provision, not on whether the conduct is of a controlling or dominating character. If the threshold were confined to coercive control, this would entail both a conduct-focused and pattern-sensitive inquiry because any such inquiry must ask whether the respondent was seeking to establish control (or dominance) over the complainant through a course of behaviour. In saying this, I borrow the definition of “abuse” under s 2(1)(c) of the VAA described as “conduct or behaviour by an individual that in any other way controls or dominates another individual and causes the other individual to fear for his or her safety or wellbeing”. Parliament’s articulation of the standard in the Second Reading of the FV Bill for the new family violence amendments points in a different direction entirely, viz, that the question is whether harmful conduct was done, and whether it caused harm to the survivor. This is a broader standard than establishing such control.

(d) Fourth, MOS Sun drew an explicit parallel between s 58B(4) and the analogous provisions in the CYPA and the VAA, where the former is intended to be aligned with the latter:

The CYPA’s definition of emotional or psychological abuse under s 5(2) covers persistent acts of rejection or degradation, threats of abandonment, isolation, confinement, and exposure to violence against related persons. These are forms of conduct that span the full breadth of emotional harm. They are not confined to controlling behaviour in the coercive control sense. I say without deciding that persistent degradation, threats of abandonment, and exposure to domestic violence are examples of emotional abuse that may occur without any overarching element of dominance. By deliberately aligning s 58B(4) with the CYPA and VAA definitions, it appears to me that Parliament signalled that the provision under the Charter was intended to be as broad as those analogues (see also below at [99]–[101]). ¹¹

(e) Fifth, the three illustrations accompanying s 58B speak for themselves. Illustration (a) states: “X spreads false rumours to third parties about X’s spouse being promiscuous. X’s spouse finds out about the rumours and is distressed.” This describes a form of reputational attack and emotional humiliation. It does not involve any element of control or dominance per se. It is not a pattern of coercive behaviour. It may even be a single act. Parliament nonetheless expressly identified it as a paradigm example of emotional or psychological abuse. If coercive control were the threshold, Illustration (a) would not belong in the Bill. Its presence is itself indicative that a broader reading of the provision is required.

(f) Sixth, both the CYPA and the VAA to which the new definition of emotional or psychological abuse under the Charter was intended to be aligned do not invoke coercive control as the conceptual thread underlying emotional or psychological abuse (see below at [99]–[101]).

94 From the above, I do not think that Parliament intended to enact a ground of coercive control as the basis for protection. It enacted a provision in terms of tormenting, intimidating, harassing, and distressing a person. These are terms that capture the full spectrum of emotional harm, of which coercive control is one serious but non-exhaustive manifestation. The approach in XPX v XPW which does not tie the inquiry to any elevated standard of coercive control but focuses on whether each limb is independently made out on the facts, is in my respectful view the sounder and more faithful reading of the statute. The assessment of seriousness operates at two levels and is adequate to weed out unmeritorious claims without the need for the importation of a test of coercive control. First, through the contextualised objective standard applied at each limb, which excludes conduct that is merely unpleasant, thoughtless, or hurtful in the ordinary sense. And second, through the necessity assessment, which independently requires the court to be satisfied that a PPO is warranted for the applicant’s protection or personal safety.

95 What, then, is emotional or psychological abuse? The question is one that this judgment has approached from several angles, ie, through statutory text, parliamentary materials, comparative survey, and engagement with the emerging body of cases in the Family Courts.

96 The answer the law provides is necessarily at broad level of abstraction. The court does not define emotional or psychological abuse with exacting precision. What it does is apply a standard, viz, the Contextualised Objective Approach as described above (at [46(c)]), as against the statutory wording (or substantive tests) under Limbs (a) and (b) to determine whether the conduct in question crosses the threshold. In interpreting these words, precedents dealing with harassment, intimidation, distress, or torment be it under the previous family violence regime (eg, those relating to “continual harassment”) or under analogous provisions in the Protection from Harassment Act 2014 (2020 Rev Ed) (eg, as to what amounts to “harassment” or even “distress” under s 3(1) of that Act) may be of assistance in illuminating the meaning of these words. A court may draw on such precedents with appropriate caution as to the precise statutory wording of s 58B(4) of the Charter. Emotional and psychological abuse is a concept whose character is inherently relational and contextual. It is concerned with the human experience of harm within intimate relationships. These are harm that, as Parliament recognised, does not reduce neatly to fixed categories or enumerable features except by way of illustration. A Family Court asked to apply s 58B(4) should endeavour as best as it can to remain anchored in the statutory language and the facts before it, rather than in an elaborated taxonomy of what abuse must or must not look like.

97 The present case is itself a modest illustration of this difficulty. The Respondent, in his closing submissions, characterised the Applicant’s conduct as a form of “emotional manipulation” directed at him. This is a characterisation I make no finding upon at this stage, and which I address in the context of my factual findings below. I raise it only to observe that in this very case, each party perceives the other’s conduct through a lens shaped by their own experience of the relationship. The question of whose perception the law should credit, and by what standard, is not a question that any abstract taxonomy can answer in advance. It is answered only by a grounded assessment of the evidence in its full relational context, which is precisely what the Contextualised Objective Approach and the well-worn principle of incrementalism require.

98 I say so also with some deliberateness, and also to build upon my point above as to the inherently human nature of the inquiry itself. The social sciences, which have studied emotional and psychological harm in intimate relationships for decades, have not produced a universally agreed definition of abuse. Researchers continue to debate where the boundary lies between harmful but normal relational conflict and conduct that rises to the level of abuse. The inquiry is inherently relational and contextual in a way that resists universal rules. The dynamics of spousal abuse may raise considerations wholly different from those of parent-child abuse, or abuse between siblings or step-siblings. The presence of an imbalance in authority as between a parent and a child, the dependency of a young child on its caregiver, the particular vulnerability of a pregnant spouse or a new mother, are but some of the considerations that differ across relationships. A rule formulated for one context may produce unjust results in another. Words or conduct that would constitute torment or intimidation in one relationship may not do so in another. The law must attend to these differences.

99 I develop this point through the CYPA and VAA. These are statutes instructive for the application of s 58B(4) considering Parliament’s express intention to align the latter with the former two.

(a) The non-exhaustive list of circumstances amounting to emotional or psychological abuse of a child under s 5(2) of the CYPA as introduced by the Children and Young Persons (Amendment) Act 2019 (Act 30 of 2019) for the purpose of assessing if a child is in need of care and protection includes: persistent rejection or degradation that are harmful to their wellbeing or sense of self-worth (s 5(2)(a)); threats to physically hurt, kill, or abandon the child (or young person, but I use the descriptor “child” for ease of convenience) or placement of child in a dangerous situation (s 5(2)(b)); isolation of the child by consistently denying opportunities for the child to communicate with any person (s 5(2)(c)); confinement in a small space for purpose of disciplining (s 5(2)(d)); influencing the child to develop self-destructive, anti-social, criminal, deviant or other maladaptive behaviour (s 5(2)(e)); and exposure (or placing at risk of exposure) to violence against related person of the child (s 5(2)(f)).

(b) The VAA’s illustrations of emotional or psychological abuse under that Act similarly include threats directed at an individual’s grandchild; and a caregiver son who is drunk, shouts and destroys belongings of an individual with dementia in his presence: see Illustrations (a) and (b) to s 2(1) of the VAA.

100 These indications of emotional or psychological abuse in the CYPA and VAA do not share a single unifying principle beyond the broad abstraction that Parliament has provided, ie, emotional or psychological abuse. They appear to span single acts and courses of conduct, direct and indirect harm, and conduct that is violent in character and conduct that is not.

101 Significantly, neither the relevant interpretative materials for both the CYPA amendments nor the VAA invoked coercive control as the conceptual thread underlying emotional or psychological abuse.

(a) On one hand, the speech of then-Minister for Social and Family Development Desmond Lee during the Second Reading of the Children and Young Persons (Amendment) Bill (Bill No 22/2019) explicitly grounded the definition of emotional or psychological abuse there in established child protection guidelines such as the American Professional Society on the Abuse of Children: Singapore Parl Debates; Vol 94, Sitting No 110; [3 September 2019] (Desmond Lee, Minister for Social and Family Development). No mention was made on the concept of coercive control.

(b) On the other, the VAA makes it plainer when it treats conduct that “controls or dominates” another as a separate head of abuse from emotional or psychological abuse. This suggests that the two are distinct categories, not one a precondition for the other. What further illustrates the difficulty of any unified taxonomy is that identical conduct may have entirely different significance depending on the relationship in which it occurs. A parent telling a child she is dirty and unloved is an act of degradation that, against the backdrop of parental authority and a child’s utter dependency, may cause profound and lasting harm. The same words spoken between adult siblings, or between spouses, may carry entirely different weight. This is not because the words have changed. It is the specific relationship that gives the conduct its power and its harm.

102 The appropriate approach, as with the cases that fall to be decided under our courts of common law, is incrementalism. This is, after all, the method the common law has always employed in developing new areas of law. Each case decided on its facts, with careful attention to the specific relationship before the court, the specific conduct complained of, and the specific harm alleged, builds the body of authority through which broader principles may, in time, be abstracted. A finding of family violence is a serious legal pronouncement. It is the law’s determination that certain conduct is unacceptable and will be sanctioned. Given the weight of that finding, there are strong reasons to require that it rest on a careful and fact-specific assessment by a judge who has considered the full subtleties and nuances of the evidence before the court, rather than on the application of any pre-formed checklist. The observations in these grounds are not intended to foreclose the development of broader principles by a future court as the case law develops over time. When the body of case law has built sufficiently through incremental decision-making, broader principles may well become capable of abstraction. We are yet in the early days of the law under the amended Charter.

103 For ease of reference, the legal framework as discussed above may be summarised as follows.

104 Section 58B(4) of the Charter defines emotional or psychological abuse through two complementary and disjunctive limbs, referred to in this judgment as Limb (a) and Limb (b) (see above at [17]–[18]). Either limb may independently establish emotional or psychological abuse, and a finding under one does not require engagement with the other. The two limbs focus on different aspects of emotional or psychological abuse:

(a) Limb (a) (“torments, intimidates, harasses or distresses a person”) focuses on the nature of the respondent’s conduct (see above at [17(a)] and especially, [51]).

(b) Limb (b) (“causes or may reasonably be expected to cause mental harm to a person, including thoughts of suicide or inflicting self-harm”) focuses on the consequences of the respondent’s conduct in terms of the mental harm (see above at [17(b)] and [51]).

105 The standard for Limb (a) is the Contextualised Objective Approach. The court asks whether the conduct was, assessed in its full relational and situational context, of a character that would objectively torment, intimidate, harass, or distress a person in the complainant’s position, having regard to the specific nature and history of the relationship, the domestic setting, any known vulnerability of the complainant, the power dynamics between the parties, and the cumulative effect of prior incidents (see above at [46(c)]). This is not a checklist but an evaluative standard (see above at [46(c)]). A purely subjective approach is rejected because it provides no objective notice of prohibited conduct and would collapse the distinction between Limbs (a) and (b) (see above at [49]–[51]). A purely decontextualised objective approach is equally untenable because it strips away the relational context that gives emotional abuse its character (see above at [52]).

106 Limb (b) contains two disjunctive components, each raising a different question and governed by a different standard (see above at [60]):

(a) The first component (“causes mental harm”) is a backward-looking question of fact. It asks: did the respondent’s conduct actually cause this complainant mental harm? The court determines on the balance of probabilities, on all the evidence, whether the harm occurred and whether the respondent’s conduct caused it (see above at [64]–[67]). The Contextualised Objective Approach does not govern this component in the same way as Limb (a). The court is finding what actually happened, not what a reasonable person would have experienced (see above at [64]–[65]). The safeguard against self-reporting concerns lies in the demanding seriousness threshold in that the mental harm required must be of a character approaching the gravity of the illustrative examples of thoughts of suicide and self-infliction of harm expressly provided in the provision (see above at [65(b)] and especially, [67]).

(b) The second component (“may reasonably be expected to cause mental harm”) is a forward-looking inquiry governed by the Contextualised Objective Approach (see above at [74]–[75]). It asks: would a reasonable person in the complainant’s actual position be expected to suffer serious mental harm of the kind described in Limb (b)? Proof of actual harm is not required (see above at [74]). This component protects complainants who may not yet suffered actual mental harm but in respect of whom the risk is objectively established (see above at [75]). This is not to say that a complainant who has suffered actual mental harm cannot rely on the second component. In fact, proof of a significant level of mental harm (be it of the types under Limb (a) such as intimidation, torment, distress or harassment or the more serious type of mental harm such as thoughts of suicide or infliction of self-harm under Limb (b)) may bolster the complainant’s case under this component (see above at [80]).

107 The following general observations apply to both limbs:

(a) The word “abuse” in s 58B(4) imports an implied threshold. Not every act causing unhappiness, frustration, or temporary distress constitutes emotional or psychological abuse (see above at [84]–[85]). The conduct must exceed the ordinary friction and unhappiness inherent in any close relationship.

(b) Coercive control is not the legal threshold for s 58B(4) (see above at [89]–[94]). It is one manifestation of the harm the provision targets, but not a precondition for its engagement. Where coercive control is established, it is relevant as context that informs findings under either limbs, as evidence explaining why apparently ordinary conduct carries greater significance in its relational setting, and to the necessity assessment (see above at [92]).

(c) Abuse may take the form of a single instance of conduct or behaviour under s 58B(5) of the Charter (see above at [12] and [18]). Where there are multiple incidents, each must be assessed both individually and as part of the cumulative pattern of which it forms a part (see above at [81]–[83]). The perpetrator’s intent is not a necessary element of a finding of family violence under s 58B(4), though it remains relevant to evaluating the nature, purpose, and context of the conduct and to the necessity assessment (see above at [77]–[78]).

(d) The ambit of emotional or psychological abuse may not be readily capable of precise prescription beyond what Parliament has provided in s 58B(4). It is to be developed incrementally through careful, fact-specific adjudication, with each case decided on its own evidence and relational context (see above at [95]–[102]).

108 I will be briefer in this part of the analysis as to the question of necessity. It is trite that once family violence is established, the court must determine whether the PPO is necessary for the “protection or personal safety” of the Applicant.

109 The necessity inquiry is fundamentally a risk assessment. It asks whether there is a sufficiently real risk that the Respondent will commit future family violence against the Applicant: XFL v XFM [2024] SGFC 103 (“XFL v XFM”) at [19]–[23].

110 The amended provision specifies “protection” and “personal safety” as distinct grounds, reflecting Parliament’s intent to broaden the circumstances in which a PPO may be ordered. “Personal safety” has a broader scope than “protection,” extending to an individual’s environment and living circumstances as they bear on wellbeing: XNY v XNZ ([16] supra) at [39]–[44]. A PPO serves a deterrent function that the ordinary criminal law may not, because breach of a PPO is itself an arrestable offence under s 63D of the Charter with the punishments stipulated in s 63C of the Charter: see XFL v XFM at [23], which laid down the same consideration, albeit under the pre-amendment family violence regime.

111 The decision of XFL v XFM at [19]–[23] is one of the more frequently cited decisions in PPO trials and understandably so given its comprehensive account of the broad factors that a Court may consider in this risk assessment. I can do no better than to reproduce the relevant parts of the learned DJ’s decision in full:

112 No single factor is determinative. The court weighs them in the round against all the circumstances. ¹² Courts will not grant a PPO in every case where family violence is established: XZU v XZV ([16] supra) at [50]–[51]. The necessity inquiry is a weighty analysis that a Court must undertake and which is one that serves to prevent unnecessary judicial intervention into family life, which in the ordinary case, should be private.

113 Based on the legal principles set out above, I now turn to apply them to the facts. Three incidents are before me. I deal with them in the order of the most recent and most serious first. In applying the Contextualised Objective Approach where relevant, I ask not only whether the Applicant herself was in fact intimidated and distressed (though I find that she was) but whether a reasonable person in her position, in this home, at this point in this marriage, in these circumstances, would have experienced what she experienced. The Contextualised Objective Approach asks precisely this: it contextualises the reasonable person by reference to everything that gives the conduct its meaning in this particular relationship. I will also explain why the conduct falls within Limb (a) rather than Limb (b), and why it exceeds the bounds of merely offensive conduct.

114 The parties do not dispute the central fact for this incident. The Respondent admits, in his affidavit at paragraph 12, that he “used my foot to push the Applicant away.” In cross-examination, when the Applicant asked him about the September incident, the Respondent did not deny the act of physical contact. His defence is not that no contact occurred, but that it was defensive rather than aggressive in character.

115 The Applicant’s account is as follows. On 30 September 2025, [A] had been crying inconsolably for over an hour in a warm room with the air-conditioning off. She asked the Respondent to turn it on. He did. She left to prepare formula. On returning, she found the air-conditioning had been switched off again. When she asked whether he had turned it off, he remained silent. She decided to bring [A] to the condominium lobby. This was a method she testified she had used since [A] was approximately two to three months old to manage the heat. She put on a baby carrier, re-entered the bedroom, and reached out to take [A]. He dodged her attempts three times. She waited approximately one minute to see if he could soothe the still-crying child. She said: “I don’t want to always involve your mum in this.” ¹³ She reached for [A] one more time. At that point, the Respondent kicked her. She described the force as “abrupt” and demonstrated it in court. This appears to me, having observed the Applicant’s portrayal of it at trial, to be a gesture akin to sharply pushing a chair away. She felt her body move despite being confined between the bed and a table. She suffered no injuries because the baby carrier absorbed the impact.

116 The Respondent’s account is that he was caring for a sick and wheezing infant, that the Applicant entered and tried to take [A] from his arms despite his protests, and that he used his foot to push her away to prevent her from snatching the child.

117 The critical fact is not in dispute. Two pieces of evidence are dispositive in my mind. The Respondent’s own affidavit states at paragraph 12 that “I used my foot to push the Applicant away to defend our son and myself from her.” This is an express admission of deliberate physical contact. Significantly, the Respondent’s own Telegram message to the Applicant provided what appeared to be an unprompted written admission of the act itself and of its effect on the Applicant. It was sent after the incident. I gave this considerable weight.

118 The Respondent attempted at trial to explain away the apology as an attempt to salvage the marriage rather than an acknowledgment of wrongdoing. I did not accept this explanation. An apology that plainly says “sorry for kicking you and making you feel unsafe and in danger” is an apology for the act and its consequences. They are the words of a man who knew he had crossed a line.

119 Further, I placed weight on Dr [B]’s letter of 15 January 2026, provided by the Respondent’s own clinical psychologist and introduced by the Respondent himself into evidence at Tab 3 of his Affidavit of Evidence-in-Chief (“Respondent’s AEIC”). This letter was prepared at the request of the Respondent for the “purpose of documenting his ongoing, voluntary engagement in psychotherapy, focused on his relationship with anger”. Most materially for present purposes, the letter goes onto state at paragraph 3 that “[the Respondent] disclosed an incident involving physical contact with his spouse, specifically an act of kicking” and that “physical aggression was identified in therapy as unacceptable behaviour and addressed as a focus for therapeutic work.” The fact that Dr [B] had specifically described the Respondent’s action as amounting to a kick is telling. This account could have only come from the Respondent. That said, I am mindful that the letter at paragraph 4 expressly disclaims any factual determination and makes clear it does not assess risk of recurrence. But what it does establish is that the kicking did in fact occur and was recognised by the Respondent at the very least as a form of physical aggression requiring further therapeutic work on his part. A respondent cannot present his clinical psychologist’s letter as evidence that no family violence occurred when that same letter records that kicking his spouse was addressed as “unacceptable behaviour” in therapy.

120 In any event, under cross-examination, the Respondent accepted that he did kick the Applicant but that this was because the Applicant did not allow him to take [A]. ¹⁴

121 A kick directed at another falls squarely within s 58B(2)(a)(i) of the Charter. Physical abuse is established where conduct “causes, or threatens to cause, personal injury or physical pain.” A kick forceful enough to move a person’s body in a confined space between a bed and a table, ¹⁵ as the Applicant described and demonstrated in court, at minimum threatens personal injury or physical pain. The Applicant when queried by me, testified that she did not suffer any bruises. ¹⁶ However, that no injury resulted is attributable, in my judgment, to the fortuity of the force applied and not due to any insufficiency of force. The Applicant’s evidence on the force used was coherent and specific. It went unchallenged in any meaningful way during cross-examination. The Respondent’s own cross-examination of the Applicant focused on whether she was trying to take [A] from him, not on whether the force he used was trivial. I was thus driven to find that there was physical abuse.

122 The Respondent’s case on self-defence, at its core, was this. [A] was unwell with a flu and was wheezing. The Respondent had turned off the air-conditioning in the belief that this would help soothe the child. The Applicant disagreed and wanted to bring [A] to the condominium lobby downstairs where there was air-conditioning. The Respondent refused, believing that exposing [A] to air-conditioning would worsen his condition. When the Applicant returned to the bedroom with a baby carrier and repeatedly reached for [A], the Respondent used his foot to push her away. He characterised this as a defensive act to protect [A] from being taken and to defend himself. I did not consider the self-defence exception to be available to him.

123 The Applicant was not posing any genuine threat, either to [A] or to the Respondent. She was reaching out to take her infant son. Even on the Respondent’s own account, she was attempting to take [A], not to harm him. Crucially, under cross-examination by the Applicant, the Respondent admitted that she did not use force in terms of strength but only that she “kept reaching out to try to grab him from my hand.” A mother persistently reaching for her own child, without physical force, cannot constitute a genuine threat to the child or to the Respondent. The disagreement between the parties was about where [A] should be taken, not about his safety. The self-defence exception in s 58B(2)(c)(i) is preventive in character and is available only where force is used to avert a genuine threat: TEK v TEJ ([13] supra) at [14]–[15]; XBC v XBD ([13] supra) at [11]–[12]. There was no genuine threat here. There were also other options plainly available such as placing [A] on the bed, seeking the mother-in-law’s help, or calling a pause. None of these the Respondent attempted.

124 On the Respondent’s own account, he was holding [A] with both hands throughout. The kick did not protect [A] from any physical danger. It drove the Applicant away from a disagreement that is fundamentally about a parenting choice, ie, about where to take her crying child.

125 Further and even accepting for argument’s sake that some justification for physical retaliation existed because of his genuine concern that [A]’s wheezing may be worsened by the Applicant exposing [A] to the air conditioning downstairs, the force deployed was disproportionate to any threat that a mother reaching for her own child without physical force could represent. The kick was forceful enough to move the Applicant’s body despite her being confined between a bed and a table. There is no proportion between that level of force and the conduct it was supposedly responding to.

126 Finally, under cross-examination the Applicant asked the Respondent what his mother’s first words were when she entered the bedroom. The Respondent stated they were: “Is [A] okay?” She did not ask whether the Respondent was alright, or whether he had managed to protect the child from the Applicant. Her instinctive question was whether [A] was safe. I found this to be the most probative window onto how the situation appeared to someone who knew all the parties and had no stake in either account. Had the Respondent merely been defending [A] from a genuine threat, the natural reaction of a concerned mother would have been to ask whether the Respondent was alright. Her question implies that the situation appeared to her as one in which [A] may himself have been at risk following and as a result of the Respondent’s actions. That is inconsistent with the Respondent’s case that he had acted to protect [A], and it reinforces my conclusion that the self-defence exception is not made out.

127 The background to the phone-snatching incident is not substantially in dispute. On 6 May 2025, [A] was slightly less than one month old at the time as he was born only in April 2025. The parties had a disagreement about the correct water temperature for preparing [A]’s formula milk. The Applicant’s position was that warm water should be used. In contrast, the Respondent’s position was that room temperature water was correct and he had produced the manufacturer’s FAQ page in support thereof (see Respondent’s AEIC at Tab 4). On the evening of 6 May, the Applicant reminded the Respondent to use warm water. At some point, the Respondent placed [A], who had been sleeping, in the bedroom. The parties were alone in the apartment.

128 Beyond this, their accounts diverge completely and this is where I have to direct my attention on the findings of fact.

(a) On the Applicant’s account according to her statement submitted to court at paragraph 6, ¹⁷ the Respondent returned to the living room in a state of anger and, without warning, kicked the footrest of an armchair aside, snatched her mobile phone from her hand and threw it onto the sofa, and then shouted at her: “You enough or not? I already told you warm water is room temperature and even found the definition. You only want to listen to what your friends say.” The shouting woke [A], who began crying. Both parties went to the bedroom. The Respondent reached [A] first and refused to pass the child to the Applicant despite her asking him to do so. She was so concerned that she messaged her mother-in-law for help before leaving the apartment for the condominium lobby. When she returned, she asked the Respondent whether he could commit to stopping violence within the family, and he replied: “I cannot guarantee that.”

(b) On the Respondent’s account in his AEIC at paragraphs 30 to 33, the parties had a verbal argument as couples sometimes do, but he denied behaving violently at any point. He did not recall the Applicant asking him not to hurt [A], denied making the “I cannot guarantee” statement, and attributed [A]’s crying to the child needing attention rather than to any outburst on his part. He further averred that the Applicant had no hesitation leaving [A] in his sole care when she left the apartment, which he said was inconsistent with any fear that he might harm the child.

129 I accept the Applicant’s factual account of this incident in full. I make the following observations.

(a) First, the Respondent’s general denial. In cross-examination, the Respondent was asked nothing about whether he kicked the footrest, snatched the phone, or threw it. His affidavit addresses the formula dispute and denies “violent behaviour” in conclusory terms, but contains no specific denial of the furniture-kicking, the phone-snatching, or the manner of the throw. I was mindful that in his cross-examination of the Applicant, he did not specifically challenge or put to the Applicant his contrary position that there was no violence whatsoever, or more specifically, that there was no kicking of the footrest, snatching of her phone, or even the throwing of her phone. That said, I appreciated the Respondent’s conduct of cross-examination in light of the fact that he was not legally trained. Where a complainant gives a particularised account about three discrete acts including the kick, the snatch, and the throw, and the respondent’s response is a general denial without engaging any of the specifics, I do not see how the complainant’s account can be said to be rebutted. Specificity unanswered speaks for itself.

(b) Second, the quality and consistency of the Applicant’s testimony. The Applicant’s account was specific in its detail. She described the sequence in that the footrest was first kicked, then phone snatched and thrown, then shouting, in the same order and with the same broad details set out in her affidavit, her police report, and her oral evidence. Her police report was not contemporaneous as it was filed only on 14 October 2025 and so I was mindful that the weight I can give to the police must therefore be reduced somewhat. The report describes the May incident as occurring in “June 2025”. This is a non-material discrepancy as to the date that I attribute to an honest error of recollection about the precise month rather than to fabrication, as the surrounding details are consistent across accounts. Beyond just the broad details, she gave her specific account of the details of what she perceived, ie, the controlled throwing of her phone, the force of the footrest kick as re-enacted by her to this Court and also her testimony that she felt scared by the Respondent’s controlled throwing of her phone. This is the kind of detail, the observation that restraint can be more frightening than abandon, that in my judgment one does not readily invent.

130 At this juncture, I state that having observed both parties giving evidence and having regard to the overall complexion of the Respondent’s responses throughout these proceedings I preferred the Applicant’s overall account. That said, there are some aspects of the parties’ incident which deserve closer attention.

(a) The first aspect is the forceful kicking of the footrest. This was not an accident. The Applicant described it as slightly more forceful than the kick of the September incident and demonstrated it as a deliberate kick of the footrest aside rather than towards her. To my mind, in such cases, in a living room, at eleven o’clock in the evening or thereabouts, alone with an infant, a spouse who kicks furniture with force conveys something to a person who witnesses it. And it is that physical aggression is available to him and that the next object it is directed at may very well not be furniture. Illustration (b) to s 58B is helpful on this point. A person “prone to smash furniture when angry” who thereby “causes the child to be distressed and in fear of personal injury” thereby commits not merely physical abuse but also emotional or psychological abuse. Evidently, aggression directed at objects rather than persons can function as distressing conduct, if not intimidation. My finding on this is reinforced by the broader context of the Respondent’s relationship with anger. The Respondent himself sought psychological counselling specifically “focused on his relationship with anger” according to Dr [B]’s report (Respondent’s AEIC at Tab 3, para 1), suggesting his own recognition that anger was a problematic aspect of his conduct.

Moreover, the Applicant’s statement was that the Respondent’s own mother had shared accounts of “his aggression in incidents within the family and church community growing up” (Applicant’s statement at para 6). This is an allegation made in passing in talking about the May incident, but I would state here that I do not doubt the Applicant’s account of this. The respondent has disputed this, stating in his AEIC at para 29 that “I am not aware of my mother ever sharing the family, the church community, and/or the Applicant about my alleged aggression. In any event, I disagree that I am aggressive.”

When directly confronted during cross-examination whether he recalled punching his youngest brother, [C], previously, the Respondent simply answered “No.” Tellingly, during re-examination when given the opportunity to clarify or elaborate on his testimony, including on this serious allegation, the Respondent chose not to address it further. Considering the evidence above in totality, I preferred the Applicant’s account as accurate in a very particular way, viz, that the Applicant had at the very least the awareness or perception of the Respondent’s aggressive ways and anger issues, even if I cannot make any finding that the Respondent has anger management issues as a matter of fact. This is an assessment for a professional to make.

Thus, even though the Respondent maintains that he was not aggressive in any way, his resort to the kicking of furniture as during the disagreement is entirely consistent with his history of difficulty managing anger and this, also, factored in my assessment of how a reasonable person aware of this history in the Applicant’s shoes would perceive his conduct.

(b) The second aspect is the snatching of the Applicant’s mobile phone and throwing it onto the sofa. This was on any view an intrusion into the Applicant’s personal space. When I had asked the Applicant in court how the throw was done, she demonstrated the manner of the throw as one in which the Respondent threw the phone in a controlled way with his palm facing upward and not as an uncontrolled hurl of anger. She was at pains to convey in her testimony her perception that the controlled manner of the throw made it “scarier” to her.

In my judgment, a person who retains restraint while deploying physical force communicates something qualitatively different from a person who loses control entirely. Loss of control may be frightening, but it is also explicable as momentary passion. ¹⁸ Restraint, by contrast, communicates deliberateness. Deliberateness may extend to the act chosen and its level, or more specifically, the choice as to whether a greater level of the act ought to have been employed. It is this quality, ie, the force deployed with control rather than in the heat of passion that I consider a reasonable person in the Applicant’s shoes who is aware of the Respondent’s history of his struggles with anger management would have found intimidating. A respondent in this position might reasonably say that his deliberateness ought not to be counted against him because he chose to act with restraint rather than abandon, and that a measured response is precisely what the law should encourage rather than penalise. I understand the force of this. But the Contextualised Objective Approach does not ask what the respondent intended to communicate, or whether his choice to exercise restraint reflects well or badly on him as a matter of character. It asks what a reasonable person in the complainant’s position would have experienced. The perpetrator’s intent, as I have stated above, is strictly speaking, unnecessary to the inquiry.

(c) The third aspect is the shouting. The shouting is addressed briefly in both accounts, and I do not rest my findings on it alone. But I note that it woke the sleeping infant, it followed the kicking of the furniture and the snatching of her phone. I further considered that it occurred in the confined walls of what parties called home. In isolation, a raised voice in an ordinary marital argument may not meet the threshold. This may of course properly be described to be the type of mere offence that Parliament has been clear to exclude from s 58B(4). But shouting that follows the display of force and the physical intrusion into one’s personal space cannot be considered as simply that. It must be seen within the context of this sequence.

131 Having heard both parties, I was persuaded that the Applicant had felt fear. This can be seen from her immediate act of messaging her mother-in-law for help shortly after the incident (Applicant’s statement at para 6). Applying the Contextualised Objective Approach, I asked: was the conduct, when assessed in its full relational and situational context, of a character that would objectively torment, intimidate, harass, or distress a person in the complainant’s position, having regard to the specific nature and history of the relationship, the domestic setting, any known vulnerability of the complainant, the power dynamics between the parties, and the cumulative effect of prior incidents? Taking into account the above points, I was satisfied that a reasonable person in this position would have been objectively intimidated and distressed by a spouse who kicked the furniture with force, snatched her phone and threw it in a controlled manner onto the sofa, and then shouted at her.

132 I would record that the appropriate basis for my finding of family violence on this incident is Limb (a), being conduct that intimidated and distressed the Applicant. It is not Limb (b) which requires the conduct to cause or be reasonably expected to cause mental harm “including thoughts of suicide or inflicting self-harm.” I should briefly explain why. Limb (b), as I have expressed above at [67], operates at the more serious end of the spectrum of mental harm covering cases where there are thoughts of suicide and self-infliction of harm. There is no evidence before me that the Applicant suffered or would be expected to suffer mental harm of that severity as a consequence of the May incident.

133 This incident started in the Respondent’s sister’s room on 3 February 2025, when the Applicant was seven months pregnant. According to the Applicant, she had obtained permission from the Respondent’s mother and sister to use the sister’s room (who was overseas studying at the time), intending to replicate in advance the arrangement where she would sleep in the same room as her infant in anticipation of [A]’s birth. The Respondent entered the room to ask her to move back to their shared bedroom. The Applicant’s account is that the Respondent entered and shouted at her loudly and abruptly: “You enough already or not?”, “Why must everything go your way?”, “You better clear this mess”, while she was resting with her back to the door. Her case is that the domestic helper spoke from outside the room from the kitchen the words: “[Respondent], talk nicely.” ¹⁹ She then packed her belongings intending to leave the house. In the living room, the Respondent came towards her and hugged her at the shoulders, at which point she felt unsafe and trembled.

134 The Respondent’s account is instead that even though there was a plan agreed to by the parties and the Respondent’s sister and parents for the sister’s room to be used as a nursery for [A] after his birth, the Applicant had unilaterally and prematurely moved into the sister’s room on 1 February 2025 even before [A] was born. It was in that context that he entered the room to ask the Applicant to move back to their shared room because his parents needed the space for work calls, and that he assisted her with her belongings. He does not recall the domestic helper making any comment (Respondent’s AEIC at paras 23 to 29).

135 The accounts are irreconcilable on whether shouting occurred. It seemed to me that the resolution of that factual dispute would depend entirely on preferring one party’s account over the others on a directly contested point, without corroboration on either side. The domestic helper was not called as a witness. The Respondent’s mother was not called. There was no contemporaneous record of the February incident in the police report either. The Applicant’s police report, filed in October 2025, refers to a shouting incident in February 2025 in one sentence only (“husband shouted at me very loudly during a disagreement. This caused me a lot of fear and anxiety”), without the same particularity that characterises her account of the May and September incidents. The absence of corroboration, on its own, does not preclude a finding. The court may of course accept uncorroborated testimony where it is sufficiently credible. But here, the accounts are diametrically opposed and neither account is inherently more probable than the other on the narrow question of whether shouting occurred. The course of examinations during trial was also sparse on the 3 February 2025 incident. I would only point out that neither the Respondent’s affidavit nor testimony in court squarely confronted the issue of whether he did in fact shout at the Applicant. This absence may be cause for eyebrows to be raised. Yet, for the reasons I have stated above I am not satisfied that the evidence is so clear on this. In these circumstances, I find that the burden of proof, which lies on the Applicant, is not discharged. I make no finding of family violence in respect of the 3 February 2025 incident.

136 Having found family violence in respect of two of the three incidents, with physical abuse on 30 September 2025 and emotional or psychological abuse on 6 May 2025, I turn to whether a PPO was necessary for the Applicant’s protection or personal safety.

137 I was satisfied that a PPO was necessary. The incidents revealed a cumulation of conduct across the first year of the marriage that justifies the court in restraining the Respondent’s conduct to protect the Applicant. In May 2025, within the first month of their infant’s life, the Respondent resorted to taking out his anger on an object and directly invaded Applicant’s space to intimidate her. In September 2025, four months later, he progressed to resort to not just direct physical contact but by kicking the Applicant who was at the time holding their infant son. The significance of the developments in this context cannot be overstated. Each incident, assessed individually, must also be assessed as part of the sequence of which it forms a part. An act that might carry limited weight as a first transgression carries greater weight when it forms part of a course of conduct. Here, the May and September incidents together form part of that course of conduct showing the Respondent’s increasing willingness to resort to force. Here, the progression of acts does not suggest to me a marriage in which conflict erupts episodically from mutual friction. The absence of further incidents between September 2025 and the hearing is more readily explained by the Applicant’s departure from the matrimonial home, which removed the opportunity for further confrontation.

138 It operated strongly on my decision also that the Respondent has not even begun to acknowledge the seriousness of his conduct. A respondent’s continued denial or minimisation of harm caused by his conduct is directly relevant to the necessity assessment. In XPX v XPW ([16] supra) at [89], the learned district judge held that “[a] parent who refuses to recognise that his/her conduct harms the child is more likely to repeat it.” The reasoning applies with equal force in the spousal context. A respondent who denies, minimises, or attributes responsibility for established family violence to the applicant has not demonstrated that the underlying disposition has been addressed. The fact that the underlying issue remains unresolved, the court may find that it is likely that the pattern of behaviour would continue into the future: see eg, XNY v XNZ ([16] supra) at [53].

139 Throughout these proceedings, the Respondent characterised the September kick as defensive, the May incident as ordinary marital argument, and attributed the cause of the conflict to the Applicant’s “emotional manipulation.” ²⁰ To quote his exact words in closing submissions, he addressed me in confident terms that “the assessment has been made that there is emotional manipulation involved by professional assessment.” It was plain to me that the Respondent believes that the cause of the disputes really was the Applicant. Furthermore, his reliance on his clinical psychologist’s opinion to argue that no family violence had occurred in closing submissions fundamentally mischaracterises the evidence. Dr [B]’s letter of 15 January 2026 expressly identified the kicking incident as “unacceptable behaviour” that “was identified in therapy as a focus for therapeutic work.” The Respondent’s own records undermine his position. His inability to acknowledge the impact of his conduct on the Applicant confounded me at trial and on reflection remains puzzling to me. This is a risk factor for future family violence. A respondent who does not recognise his conduct as harmful is a respondent who has not confronted the cause of it.

140 One further matter reinforces my conclusion on necessity. Dr [B]’s letter of 5 December 2025, produced by the Respondent himself, identifies “developing emotion regulation skills” as a primary objective of the therapeutic work from the outset. The Respondent’s own mother, according to the Applicant’s unchallenged (albeit strictly speaking, hearsay) evidence on this point, urged him to seek professional help because she was worried about his “outbursts” and their impact on [A]. These speak to the Respondent’s struggles with emotional dysregulation and that those closest to him recognised as requiring professional attention. A PPO would keep the Respondent aware of the need to behave civilly in any future interactions with his spouse.

141 The Respondent’s voluntary engagement in therapeutic work, while commendable, does not alone negate the necessity of a PPO. Dr [B]’s letters shows consistent voluntary engagement since October 2025, with approximately fifteen sessions by the time of hearing. But voluntary therapy without accountability provides no guarantee of the Respondent’s continued engagement with such therapeutic work or even that any behavioural change would result. This is doubly so here where the respondent continued to blame the other party for the violence. In my judgment, a PPO, with the legal consequences for breach, enforces the accountability that voluntary therapy cannot.

142 Finally, the parties’ ongoing parental responsibilities require continued interaction. They are the parents of [A], who requires both their involvement. The risk of future family violence is very much alive. It arises directly from the reality of a co-parenting relationship that will require them to communicate and manage disagreement for years to come.

143  I also note, as circumstantial evidence of the Applicant’s continued fear, the Applicant’s immediate departure from the matrimonial home with a five-month-old infant on the night of 30 September 2025, and her continued residence at her mother’s home through the date of hearing. A person who did not genuinely fear future violence would have had little reason to uproot herself and her infant so abruptly, or to remain away from the matrimonial home for months. I trust that this PPO would in some ways address this.

144 In addition to the PPO, I made a counselling order against both parties under s 60E(2)(a) read with s 60E(3)(a) of the Charter on the court’s own initiative. This power exists independently of any application by the Applicant and reflects the court’s recognition that a PPO alone, while necessary, does not address the difficulties of the marriage. Counselling may assist them in developing the communication and conflict resolution strategies that [A]’s upbringing will require of them.

145 For the foregoing reasons:

(a)  I was satisfied on a balance of probabilities that the Respondent has committed family violence against the Applicant, namely:

(i) physical abuse within the meaning of s 58B(2)(a)(i) of the Charter in respect of the 30 September 2025 incident; and

(ii) emotional or psychological abuse within the meaning of s 58B(4)(a) of the Charter in respect of the 6 May 2025 incident.

(b)  I was further satisfied that a Personal Protection Order is necessary for the protection and personal safety of the Applicant under s 60A(1)(b) of the Charter.

146 A Personal Protection Order was thus granted restraining the Respondent from committing family violence against the Applicant. A Counselling Order was further made of the court’s own motion under s 60E(2)(a) read with s 60E(3)(a) of the Charter, directing both parties to attend counselling.

147 I conclude these grounds with some remarks. Before proceeding to trial, I had taken the opportunity to explore with the parties whether an undertaking by the Respondent might provide a basis for resolving the matter without the need for a contested trial. I did so in the awareness that this was a newly forged marriage with a very young child, and that an agreed resolution, if the parties could reach one, would better serve their interests as co-parents than an order made after contested (and likely, emotionally charged) proceedings. I explained to the parties that an undertaking in the form of a formal promise recorded by the court would carry real legal weight, and that a future breach could be placed squarely before the court in any subsequent proceedings. I took care to present this as one of the available courses. I made equally clear that the Applicant was fully entitled to pursue her application to trial if she preferred. The Applicant was not persuaded to accept an undertaking. She explained that what had followed the incidents was not remorse from the Respondent but justification, with him characterising her as emotionally manipulative and that the inaccuracies she perceived in the Respondent’s affidavit had removed whatever confidence she might otherwise have had in a promise alone. She was entitled to that view and I understood her stance. For the avoidance of doubt, none of these considerations here has animated my decision and reasoning in the earlier parts of this judgment. The matter proceeded to trial.

148 I would also record that I had closed the hearing with some observations addressed to both parties. At that time, I observed that this was a new marriage. The Respondent’s conduct had been harmful, and the Applicant was entitled to the protection this order provides. But the orders made need not mark the end of the possibility of a functioning co-parenting relationship. The PPO establishes the acceptable boundaries within which that relationship must operate. The counselling order creates the opportunity for both parties to develop healthier ways of managing the inevitable conflicts of a marriage and parenthood together. What they make of that opportunity was for them to decide. [A]’s wellbeing will be shaped in no small part by whether his parents can, over time, find better ways of working through their differences.

149 The Respondent has appealed my orders to the High Court. This Court has taken the matter as far as it can.