Next Article in Journal
A Convergence of Violence: Structural Violence Experiences of K–12, Black, Disabled Males across Multiple Systems
Next Article in Special Issue
Continuity of Parental Responsibility in Child Abduction Cases: Lesson Learned from the Case of Z. v. Croatia
Previous Article in Journal / Special Issue
The Interplay between the 1980 Hague Convention on the Civil Aspects of International Child Abduction and Domestic Violence
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:

Protecting Mothers against Domestic Violence in the Context of International Child Abduction: Between Golan v Saada and Brussels II-ter EU Regulation

School of Law, Università degli Studi di Milano-Bicocca, 20126 Milano, Italy
Laws 2023, 12(5), 79;
Submission received: 2 August 2023 / Revised: 15 September 2023 / Accepted: 15 September 2023 / Published: 19 September 2023


The need to protect victims of domestic violence is becoming increasingly more important in many States. The 1980 Hague Convention on international child abduction, which in principle requires the child’s return and apparently leaves little scope for protecting the child’s mother, is at times perceived as being at odds with this need. The 2022 US Supreme Court’s judgment in Golan v Saada is set to become a leading case with regard to abductions occurring against the backdrop of domestic violence. Although the USSC, out of necessity, considers the issue from the viewpoint of the US legal system, the impact of the decision will be felt well beyond the country’s borders. This paper will start by analysing the legal arguments developed by the USSC in finding that ameliorative measures are not required by the 1980 Hague Convention, but lie at the discretion of the courts, as well as the general principles laid down by the USSC to guide the exercise of that judicial discretion. Furthermore, the rationale for—discretionary, but still relevant—protective measures will be measured against the Brussels II-ter EU Regulation, which has established a different legal framework for EU Member States. In contrast to the position under pure Hague cases, the EU Regulation now clearly calls on the courts of the State of refuge to guarantee the child’s physical and psychological safety by directly adopting provisional measures, which will apply to the child upon return to the State of habitual residence and which are recognizable and directly enforceable in that Member State. It will be argued in this paper that ameliorative/protective measures offer a means for filling a gap that is increasingly being felt within public opinion, but that could undermine the efficacy of the 1980 Hague Convention. The best way of ensuring that domestic violence cases are not neglected, while at the same time remaining within the confines of the 1980 Hague Convention, would be to adopt expeditious, substantively well-defined, and effective protective measures.

1. Introduction

USSC decisions in abduction cases are particularly relevant for the application and uniform interpretation of the 1980 Hague Convention. The impact of these decisions reaches far beyond the USA and impacts all Contracting States. To date, the Supreme Court has issued three decisions in abduction cases: Abbott v Abbott, Monasky v Taglieri, and the decision discussed in this paper Golan v Saada.1 On closer analysis, the last-mentioned decision originates from and aims to solve an internal dispute among US courts, which adopted different positions regarding protection measures and the return of an abducted child. The most immediate effect of the decision was to reverse the directions given by the Second Circuit Appellate Court in the Blondin case.2
However, aside from the internal impact within the United States, the USSC decision is of great interest also beyond its borders because of the factual situation underlying the case, one involving undisputed domestic violence. At a time when the fight against domestic violence has finally started to receive the attention it deserves, the need to protect domestic violence victims’ sometimes conflicts with the principal aim of the 1980 Hague Convention. This is the first time that a supreme court has been faced with the difficult task of balancing domestic victims’ legitimate need for protection with the aim under the Convention of securing the prompt return of the child. By stating that consideration of ameliorative measures is not required by the Convention, the Golan decision may appear to suggest a conservative approach, which would appear to be outdated. Last but not least, the Golan decision is relevant in terms of the impact it may have on the construction of similar—but not identical—EU rules.

2. The Golan v Saada Case

The facts underlying the decision arise more frequently in practice than one might think when reviewing court decisions. The case also laid bare the inefficiency and inadequacy of the law. It should be pointed out that the child involved in this story, B.A.S. Saada, was abducted and taken to US in 2018 at the age of 2 and a half. Today, after four years of legal proceedings, at the age of 7, notwithstanding the numerous (in total four) return orders issued by the US Hague court, he has still not been returned to Italy, the State where he was previously habitually resident, and a final custody decision is still to be taken. It must now be concluded—in the child’s best interests—that he will never be returned.

2.1. The Facts of the Case

Narkis Golan, a US citizen, and Isaac Saada, an Italian citizen, met in 2014 in Milan. In 2015, Golan moved to Milan in Italy, where the couple married3 and started living together. In June 2016, the couple had a child, B.A.S. The relationship appeared to be abusive from the outset. It is undisputed that Saada behaved unacceptably towards his wife. As the court found, he ‘physically, psychologically, emotionally, and verbally abused’ his wife; furthermore, he ‘called her names, slapped her, pulled her hair, threw a glass of wine in her direction, and threatened to kill her’.4 While the man did not appear to have a history of direct violence or neglect towards B.A.S., much of the acknowledged behaviour was committed in the presence of the child, thus constituting what should today be regarded as indirect violence. Mrs. Golan, who felt very isolated and was apparently also subject to control and restrictions from her parents-in-law,5 sought protection from the Italian police, who referred the matter to the Italian social services. The Italian social services launched an investigation, which resulted in a report noting the highly concerning situation and proposing that Mrs. Golan and the child be placed in a safe house, while the matter was referred to the courts. At that point, Mrs. Golan retracted her statements, declaring that she felt able to handle the situation herself and did not want to leave her home.6 However, when the opportunity arose, she ran away. In August 2018, after attending her brother’s wedding in New York, she disappeared and moved to a confidential domestic shelter. Since then, B.A.S., who is now 7 years old, has been in the USA.

2.2. The Judicial Proceedings in the District and Appellate Court (Second Circuit)

Mr. Saada promptly sought the return of his son and, in September 2018, applied for a return order before the New York District Court. The Court held that: (a) the retention was unlawful7; (b) there was a grave risk of psychological harm given the clear evidence of violence towards the mother; (c) the court was required under Second Circuit case law8 to determine whether there were any ameliorative measures or ‘undertakings’ that could minimise the risk of grave harm and allow the safe return of the child; and (d) that this was the case since Saada had agreed to the adoption of a package of measures that, taken as a whole,9 would alleviate any risk faced by B.A.S. upon return. On that basis, on 19 March 2019, the District Court ordered the return of the child (Saada I).10
On appeal, the Second Circuit vacated the District Court’s decision, finding that the ameliorative measures in question would not eliminate the grave risk of harm because they were not directly enforceable in Italy; moreover, the court had not obtained a sufficient guarantee of compliance. The Appellate Court thus instructed the District Court to consider whether the Italian court could itself issue an order prohibiting Mr. Saada from approaching Mrs. Golan or visiting B.S.A. without her consent (June 2019, Saada II).11
On remand, the District Court liaised with Italian Central Authority and instructed the parties to petition the Italian courts accordingly. In December 2019, following an application by Mr. Saada, the Court of Milan issued a far-reaching and comprehensive protection order, addressing multiple aspects relating to the protection of the child and the mother and providing support and direction for both parents (although especially for Mr. Saada), including psychological counselling and cognitive behavioural therapy. The order was meant to take effect at the time when Mrs. Golan and B.A.S. actually returned to Italy and would remain applicable for one year, with the possibility of extension thereafter.
Satisfied with this outcome, the District Court again ordered the return of the child (May 2020, Saada III).12 Mrs. Golan again appealed to the Appellate Court, which, however, again confirmed the District Court’s decision, stressing the ongoing involvement of the Italian courts, and therefore, the return of B.A.S. (October 2020, Saada IV).13
Mrs. Golan then filed a petition for a rehearing in the Second Circuit, which was denied, and a motion to vacate the District Court judgment, which was also dismissed. Having exhausted all ordinary appeals, Mrs. Golan now turned to the Supreme Court. She filed a petition for a writ of certiorari with the strong support of a number of bodies from various backgrounds. The petition was granted, and the case finally arrived before the highest court in the land. The case before the Supreme Court attracted great interest both within the legal community and amongst the wider public. No more than twelve amicus curiae briefs were filed with the court.14

2.3. The Legal Question before the USSC: Is Seeking and Crafting Ameliorative Measures Always Required by the 1980 Hague Convention, despite a Determination of Grave Risk?

The legal question put before the USSC by the petitioner, Mrs. Golan, was precisely defined and limited to a very specific issue. It is important that this be properly understood so as to rightly appreciate the scope of the decision. The relevant issue was
‘Whether, upon finding that return to the country of habitual residence places a child at grave risk, a district court is required to consider ameliorative measures that would facilitate the return of the child notwithstanding the grave risk finding’ (emphasis added).
The question was, therefore, not whether there had been any domestic violence, nor whether the return per se would expose the child to a grave risk. Moreover, there was no discussion as to whether or not the ‘package’ of protection measures envisaged by the District Court was adequate to protect the child. Although these questions had been discussed in the proceedings and had been addressed by various arguments, the issue referred to the Supreme Court did not concern either the interpretation or the assessment of any of them. The sole question was whether considering ameliorative measures was a formal requirement under the convention, or lay at the discretion of the court.
This question originated from the situation existing within the US. Courts dealing with Hague return cases are faced with conflicting case law from federal courts of appeals as to whether, after a finding that there is a grave risk of harm, a trial court must consider possible ameliorative measures to facilitate the return of the child.
On the one hand, the First, Eighth, and Eleventh Circuits have indicated that, having established a grave risk of harm, the trial court is under no obligation to consider ameliorative measures.15 In contrast, the Second, Third, and Ninth Circuits require a district court to consider—and craft when needed—a full range of ameliorative measures that would permit the return of the child in all situations where a grave risk is established.
The leading decision in this case is Blondin v Dubois, which was delivered by the Second Circuit in 1999.16 The case concerned a mother who left France with her children to escape the abusive father, who repeatedly beat her and threatened to take her and the children’s lives. The district court found that the requirement of evidence of grave risk on return was met and refused the application for return. However, the Second Circuit Appellate Court annulled that decision and remanded the case to the District Court, directing it to consider ameliorative measures that ‘might allow both the return of the children… and their protection from harm, pending a custody award in due course by a French court with proper jurisdiction’.17 The court thus referred to its own precedent in the Blondin II18 decision, which required that, where a district court finds there to be a grave risk of harm, it ‘must examine the full range of options that might make possible the safe return of the child’ before denying repatriation. The Court considered this rule to ‘honor the important treaty commitment to allow custodial determinations to be made—if at all possible—by the court of the child’s home country’.
This decision was subsequently relied upon by other Circuits,19 thus creating uncertainty and discrimination among petitioners, depending on where they were located when seeking refuge within the US. The petition to the Supreme Court challenged this approach and the resulting situation; not surprisingly, the USSC granted certiorari.
It is important to stress that the relevant question was only whether there may be a requirement to consider ameliorative measures. Even if this were to be answered in the affirmative, it would never impinge on the court’s discretion to refuse return. Whether to order return or not, and on what conditions, is always in the discretion of the court of first instance. As was pointed out by one of the few amicus curiae filings for the respondent Saada, ‘a court may always exercise its discretion to reject ameliorative measures after considering whether or not they will be effective’ (Brief for Amici Curiae 2022, p. 15). The issue at stake was by contrast only whether ‘the court is required to consider ameliorative measures even if the court ultimately decides—as it did in Simcox—that undertakings or protective measures will not be effective and refuses to order the child’s return’ (Brief for Amici Curiae 2022, pp. 16–17; emphasis added).
Once the question has been clarified this way, its answer would probably appear to be obvious in other countries. It is submitted here that the question raised is one that is highly specific to the US system because of the specific guidance given by a few appellate courts. In other Contracting States, it would probably not be disputed that a court always retains full discretion over how to handle the case (Chalas 2023).

3. Arguments of the Parties

The arguments raised by the petitioner Golan, and the amici curiae supporting her position, ran along the following lines:
(a1) A treaty should be interpreted according to its literal text. Although this argument often appears at the outset of acts, it was not consistently applied throughout the respective reasoning.
(b1) Ameliorative measures adopted in the form of undertakings are inefficient because they are not enforceable: They rely on the seriousness of the party making the undertaking and, therefore, on the parties’ respective decisions over whether or not to comply.
(c1) Ameliorative measures are highly inappropriate in cases involving men with a past of abusing women because abusive men are inherently unable to comply to court orders.
(d1) In order to protect the child and the mother, crafting appropriate protective measures requires an in-depth knowledge of the case, which is at odds with the need to decide on the Hague proceedings in the most expeditious manner.
(e1) Furthermore, ameliorative measures that are too detailed and specific conflict with the principle that the State with jurisdiction over the merits must retain competence.
The respondent Saada, and the amici curiae filed in support of his position on the return of the child argued mainly along the following lines:
(a2) When interpreting a treaty, the analysis begins ‘with the text of the treaty and the context in which the written words are used’ (Air France v Saks, 470 U.S. 392, 397 (1985)).
(b2) The requirement that a court must consider whether ameliorative measures are sufficient to protect the child’s safety is inherent within the nature of the question and is supported by the operational framework of the Convention and the accompanying Explanatory Report.
(c2) The United States should follow international practice, special commissions, and guides to good practice. In particular, courts in the United States should look to the Guide to Good Practice for Article 13(1)(b) (HCCH 2020) with regard to the assessment of ‘grave risk’.
(d2) Ameliorative measures can be effective in protecting the child, as has been shown by other international instruments.
(e2) The mandatory consideration of ameliorative measures does not require the court to adopt ameliorative measures in order to obtain the return of the child. A court always retains full discretion over whether to order the child’s return.

4. The Decision of the USSC: Ameliorative Measures Are Not Required by the Convention, and It Lies in the Court’s Discretion to Decide on Them…

The Supreme Court’s decision, which ran to 16 pages, is a very short one. The first eight pages set out the facts at the heart of the dispute whilst the remaining eight develop its legal reasoning.
Firstly, the Court clarified its approach to the interpretation of treaties by referring to the general principle that ‘interpretation of a treaty … begins with its text’. This is a long-established principle in the US legal tradition and is strongly felt by the USSC in its current composition, which is dominated by ‘textualist’ interpretations.20 This approach often lies at the heart of decisions concerning rules based on international treaties and clearly implies the rejection of any purposive (i.e., functional) or evolutive interpretation of the treaty.
Based on this understanding, the Court found that the 1980 Hague Convention imposes a general obligation to return the abducted child, but that a Hague court is ‘not bound’ to order the return of the child whenever a finding of grave risk has been made. In other words, the grave risk defence under Article 13(1)(b) ‘lifts the Convention’s return requirement, leaving a court with the discretion to grant or deny return’. ‘Discretion’ thus becomes the key to the decision, and determines its ultimate outcome.
It thus readily follows that ‘nothing in the convention’s text either forbids or requires consideration of ameliorative measures in exercising such a discretion’. Moreover, there is no doubt that ameliorative—or protective—measures are not mentioned by the Convention. Thus far, the Court’s arguing is unobjectionable and objectively sound. The following passages may, however, appear less convincing.
Responding to the arguments made by the respondent Saada, the Court analysed whether the need to put in place measures to mitigate a grave risk of harm may be considered an ‘implicit’ requirement, i.e., if the consideration of ameliorative protective measures must be assessed within and form a part of any ‘grave risk’ analysis.
In doing so, it examined the relationship between ascertaining the existence of a grave risk and the availability of ameliorative measures. The delicate question was whether these constitute two separate issues, arising consecutively, or whether they should be assessed together as one single complex but unitary issue.
This may appear to be a procedural, highly technical, question concerning only the adoption of ameliorative measures. On the contrary, given the special weight afforded to the court’s discretion following a finding of grave risk, the way in which such a relationship is construed becomes crucial (Trimmings and Momoh 2021). If the overall circumstances must be examined together, i.e., considering measures that might reduce a prospective, alleged risk before looking into the likelihood of such an allegation, this will have the effect of partly reducing the court’s discretion following a finding of grave risk. In fact, a court will not even conclude that a grave risk exists, because even if there were such a risk, it would be reduced and minimised for the purpose of returning the child. On the other hand, if the two questions are distinct and separate and the court is first required to establish whether there is a grave risk, and only then to investigate the existence of possible ameliorative measures, then there is more scope for the exercise of discretion. The need for—and efficacy of—protective measures is assessed having regard to a risk that has already been found to exist by the court.
This issue was debated at length within the Hague Conference’s Special Commission when drafting the Guide to Good Practice on Article 13(1)(b) (HCCH 2020). A first draft, released in 2017 at the Seventh Meeting of the Special Commission, offered national courts two different approaches depending on the facts of the case, national practices, and procedures (HCCH 2017). The result was seen as being complicated and confusing and the final Guide to Good Practice simplifies the position by offering a step-by-step approach. First of all, the court should consider whether the parties’ allegations could constitute a grave risk. As it is clarified, ‘Broad or general assertions are very unlikely to be sufficient’, and a unilateral statement will clearly not imply that a grave risk has been established. The second step, in fact, envisages that:
‘the court determines whether it is satisfied that the grave risk exception to the child’s return has been established by examining and evaluating the evidence presented by the person opposing the child’s return/information gathered, and by taking into account the evidence/information pertaining to protective measures available in the State of habitual residence. This means that even where the court determines that there is sufficient evidence or information demonstrating elements of potential harm or of an intolerable situation, it must nevertheless duly consider the circumstances as a whole, including whether adequate measures of protection are available or might need to be put in place to protect the child from the grave risk of such harm or intolerable situation, when evaluating whether the grave risk exception has been established.’ (…)
‘In cases where the taking parent has established circumstances involving domestic violence that would amount to a grave risk to the child, courts should consider the availability, adequacy and effectiveness of measures protecting the child.’
On the face of these passages, one might conclude that, in the view of the delegations of the Contracting States that drafted and approved by consensus the 2022 Guide to Good Practice, the existence of a grave risk is dependent and conditional on whether or not ‘adequate measures of protection’ (i.e., according to US terminology, ameliorative measures) are available.
The US Supreme Court attached very little weight to the work and conclusions of an authoritative and attentive body such as the Special Commission of the Hague Conference, with lip-service being paid to its work in a footnote,21 but clearly not allowing its work to have any impact on the decision itself. Similarly, and in contrast to other USSC opinions, there is here no room for any consideration of the practice of other Contracting States to the 1980 Hague Convention. The Supreme Court concisely affirmed that ‘the question whether there is a grave risk (…) is separate from the question whether there are ameliorative measures that could mitigate the risk’22 (emphasis added). Later on, the Court admittedly conceded that these two ‘will often overlap’ and that ‘in many instances, a court may find it appropriate to consider both questions at once’.23 However, this does not alter the previous finding on the breadth of the discretion available to the Hague court. Indeed the Supreme Court took the view that ‘the court’s discretion to determine whether to return a child (…) includes the discretion whether to consider ameliorative measures that could ensure the child’s return’.24 The point was thus quickly resolved by what appeared to be also a reproach of the Second Circuit: ‘By imposing an atextual, categorical requirement that courts must consider all possible ameliorative measures in exercising this discretion’, the rule affirmed in Blondin I in practice ‘rewrites the treaty’.25

5. …. and the Guiding Principles for Exercising Such Discretion

While the above-mentioned conclusion might appear to have resolved the controversial question raised before the USSC, it did not bring the case as a whole to an end. Having acknowledged the broader scope for the court’s discretion, the USSC also sought to provide guidance to direct the exercise of that discretion. This is, in the present author’s view, the most important part of the decision. It must be noted, however, that strict adherence to the text of the Convention, as proclaimed at the beginning of the decision, is less evident in this second part. In fact, without mentioning it, the court appears to follow a purposive—or functional—interpretation, i.e., an interpretation of the rules that is based on the aims of the convention.
After noting that the text does not contain any reference to ameliorative measures, and having, therefore, drawn the obvious conclusion that there is ‘no obligation to consider’—and still less to order—ameliorative measures, the court went on to hold that a court ‘ordinarily should address ameliorative measures raised by the parties or obviously suggested by the circumstances of the case, such as in the example of the localized epidemic’26 (this example was raised by the court itself a few pages before, as one where the risk could be minimised by returning the child to a different part of the country).
The lack of any reference to ameliorative measures within the text of the Convention, which was emphasised in the first part of the decision, was now supplemented and completed by a purposive interpretation of the Convention, which led to the conclusion that these measures—though not mentioned at all—should nonetheless ordinarily be addressed whenever they are appropriate in the given case.
The added value (but also the creative part) of the decision is to be found in section B of the decision, where the USSC elaborates on three ‘legal principles’ that should guide courts when considering/ordering protective measures. In doing so, the USSC clearly departed from a textual interpretation in favour of a more teleological or purposive one.
The purposive approach is first apparent when the USSC identifies the core aim of the Convention, which is found to lie in the ‘protection of the interests of children and their parents’. The Court also stated that return must not be pursued at all costs, again reproaching the Second Circuit for having ‘improperly elevated return above the convention objectives’.
This is a worthy statement, although it would probably have needed a few more sentences to be properly appraised. While its content is essentially unobjectionable, it is submitted here that this conclusion does not flow directly from the literal wording of the Convention but is inferred from its current application and understanding, especially as clarified and developed by the case law of the European Court of Human Rights after the Neulinger and X decisions (among many: Beaumont et al. 2015; Mc Eleavy 2015).
Moreover, one of the difficulties in applying the 1980 Hague Convention lies in striking the right balance between the interests of children as a category, which is mainly served by their return after abduction, and the interests of the individual abducted child in a given case, which sometimes may be served by not allowing return to the State of habitual residence but rather remaining in the State of refuge. In this regard, to refer to the ‘protection of the interests of children’ does not really add much clarity or guidance but may, instead, cut both ways. Furthermore, it is submitted here that the juxtaposition between protecting the interests of a child on the one hand and ordering his/her return on the other hand may entail a risk of undermining the structure of and procedures under the 1980 Hague Convention.
Finally, to refer to the (best) interests of the child may appear to be injudicious in view of the widely experienced difficulties in convincing courts seised with return proceedings that they should not carry out a proper welfare-interests test before returning the child; that such a test should be limited to the reasons for refusing return; and that return to the State of previous habitual residence is in principle in the interest of any abducted child. In sum, as is often the case, the reference to ‘best interests’ is inherently meaningless, given its indeterminacy.
In practice, the three legal principles offered by the Supreme Court as a guide to the exercise of discretion by courts when considering the need for ameliorative measures are more useful.
The first refers to the need to ‘prioritize the child’s physical and psychological safety’.27 There is no question that the child’s safety is paramount. Indeed, this is included in the idea that the court should (but need not necessarily) refuse return when it is convinced that there is a grave risk, and that grave risk cannot be minimised so as to be acceptable. Obviously enough, if the ameliorative measures proposed are ineffective or if the Court is convinced that they would not succeed in minimising the risk, then the child should not be returned.
The judgment then provides some examples of when such a risk can be identified. The following passage is so important that it is worth quoting in full.28
‘Sexual abuse of a child is one example of an intolerable situation. See 51 Fed. Reg. 10510. Other physical or psychological abuse, serious neglect, and domestic violence in the home may also constitute an obvious grave risk to the child’s safety that could not readily be ameliorated. A court may also decline to con- sider imposing ameliorative measures where it reasonably expects that they will not be followed. See, e.g., Walsh v Walsh, 221 F. 3d 204, 221 (CA1 2000) (providing an example of a parent with a history of violating court orders)’.
From a methodological point of view, it is worth noting that all of these examples are derived from internal US practice, with no reference being made either to the Perez Vera Report or to the HCCH Guide to Good Practice (HCCH 2017), the principal aim of which was specifically to give Contracting States some directions on how to assess grave risk.29 In this regard, it is a pity that one of the relevant actors on the Special Commission, and indeed one that is looked upon by other Contracting States in a search for uniformity, has missed the opportunity to support this relevant exercise of international harmony.
The second requirement is that protective measures should not impinge upon the assessment of the merits of the case and encroach upon the competence of the courts of the State of habitual residence. This limitation stems from Article 16 of the 1980 Hague Convention, which posits a clear division of competences between the Hague return proceedings and the custody proceedings to be held in the State of habitual residence. While this clarification certainly makes sense in terms of global guidance, the issue seems unrelated to the Golan case, where the prospective protective measures were actually adopted by the Italian court that would have been competent over the decision on custody. However, in past US practice, the courts have at times elaborated measures that have ended up being too intrusive on the substance of custody, access, or maintenance decisions, and the US Supreme Court rightly took the opportunity to clarify the position.30
The third requirement, which refers to the need that consideration of ameliorative measures must not affect the expeditious nature of return proceedings was apparently more relevant to the case at hand. There is no need to emphasise here the importance of expedited proceedings when deciding the return of an abducted child. Indeed, the entire rationale of returning a child to his/her previous State of habitual residence presupposes that both the return order and the actual return will take place within a short space of time. If the child is returned after a considerable length of time, his/her return will entail a new uprooting from the place where the child was forcibly and painfully integrated.
While the principle is undisputed, its practical implementation is often overlooked and disregarded. It is clear from the return proceedings before the District Court how difficult it is to craft and put in place protective measures for the return of the child, especially because these measures need to be effective and enforced elsewhere, i.e., in the State of the child’s habitual residence. The situation is especially complex in jurisdictions, such as the US, that are not a party to any international judicial cooperation instruments that may help in ensuring the enforcement of these measures.31
The Golan case provides a good example of this. Because the Second Circuit had reversed the first decision of the District Court on the grounds that there were ‘insufficient guarantees of performance’, on remand, the District Court carried out an extensive examination to ensure that any ameliorative measure put in place would be truly effective. The District Court, therefore, turned to Italian courts—i.e., the courts of the State of the child’s previous habitual residence—to ensure that this was the case. An Italian protective order was, indeed, quickly sought (in three months: from September to December 2019). However, another six months elapsed (until June 2020), resulting in a total of nine months, before the (second) decision on return was taken, again by the District Court. It should, however, be noted that, absent any judicial cooperation treaty, if safe return is to be guaranteed, there is no real alternative to seeking protective orders in the State of habitual residence. Considering the way in which the proceedings as a whole developed, it seems that their exceptional length was not caused by the time necessary to obtain an effective protective order in the State where it was actually needed. In order to avoid unnecessary delays, better case management could have been pursued at other stages, including during the final stage when the USSC decided to remand the case for the third time to the (same) District Court in New York.
While all of these directions are useful and relevant, one is struck by the fact that, in a case characterised by undisputed serious violence, the USSC offered only minor guidance on how to handle cases involving domestic violence. The legal reasoning on ameliorative measures is carried out with an eye on the general rule on grave risk and includes only a minor reference, amongst other hypotheses—such as the one quoted above in this paragraph—the impact of domestic violence on the child. This is probably in line with the approach embraced by the USSC in following the ‘original’ interpretation of the convention. As the 1980 Convention does not refer to such a scenario, the USSC deemed it better to develop its reasoning with regard to the general rule on grave risk. However, a number of authoritative, in particular US-based, legal scholars have long since argued that domestic violence cases deserve a special consideration and should be made the object of special attention (Weiner 2021; Trimmings et al. 2023). Given the strong and lively accusations recently levelled, objecting at how the 1980 Hague Convention is incapable of protecting abducting mothers fleeing from domestic violence, coupled with the initiatives taken to question the Convention as a whole or to establish its inadequacy (Barnett 202332; Trimmings et al. 2023; Weiner 2021; Pahrand 2017), this would have been an ideal opportunity for the USSC—which is a major player in the international arena—to signal its attention and consideration for a social need that can no longer be ignored.

6. The Aftermath of the Decision: The Death of Mrs. Golan and the (Still Pending) Provisional Custody Proceedings Involving B.A.S. in the US

On remand from the USSC, on 31 August 2022, the District Court issued, for the third time, an order directing the return of B.A.S. to Italy, subject to the conditions set out in that court’s original opinion (including the Italian protective order, which the court had ordered the parties to negotiate and which the Italian courts had issued). Mrs. Golan again appealed (for the fourth time) to the Second Circuit Court of Appeals.
Mrs. Golan argued before the Court of Appeals that the District Court abused its discretion by issuing the return order without taking evidence concerning a number of factors that the Supreme Court opinion indicated should be weighed when applying the Convention. These factors included the safety and well-being of her son (who was 6 years old at the time) if he were required to be returned to Italy despite the finding concerning grave risk and other developments, including the fact that the child was on the autistic spectrum. Mrs. Golan also asked the Court of Appeals to dismiss the case and to allow B.A.S. to remain in the United States pending the resolution of the custody issues, which the parents would address in separate proceedings.
On 19 October 2022, whilst these appellate proceedings were pending, Narkis Golan suddenly passed away. She was 32 years old; B.A.S. was 6 years old at the time (Fidler 2023).
Following Narkis Golan’s death, her sister, Morin Golan, took over physical custody of six-year-old B.A.S. and filed a family offense and custody action in the New York State Kings County Family Court. On 20 October 2022, the Family Court issued an ex parte order directing that the child ‘not … be removed from the care of Morin Golan until further court order’, that B.A.S.’s passports be held by Ms. Golan and that the child not be removed from the jurisdiction of the Kings County Family Court. The Family Court also issued a temporary order of protection, directing the father to ‘stay away’ from B.A.S. and the aunt, ‘[s]ubject to the order of supervised visitation of the Tribunal of Milan.’ The Children’s Law Center (a charity based in New York) was further assigned to represent B.A.S. interests.
Arguing that the mother’s death represented a substantial change of circumstances for B.A.S., Morin Golan and the guardian for B.A.S. submitted that the NY District Court consider the new and unexpected child safety concerns, and refuse his return. The Italian provisional measures were meant to take effect upon the return of B.A.S. with the mother. If the child would be returned alone, this would raise entirely unexpected and unaddressed new concerns.33 These included determining where the child was to be placed; dealing with the difficulties he would experience in relating with a father he had not met in the last 4 years; settling in a country of which he had no memory and whose language he did not speak; and difficulties in adapting to a new situation, including his autistic spectrum disorder.
It is difficult to believe that returning the child after 4 years, and under the circumstances described, would be in his best interests. At the moment of the writing the (provisional) custody case is still pending, as are the Hague return proceedings. In the personal opinion of the present author, whatever the right decision would have been at an earlier stage, to return B.A.S to Italy now would represent a grave and substantial failure of the law.

7. The Golan Decision Seen from the EU and Its Impact on the (Recast) EU Legal System

As mentioned above, the Golan v Saada opinion sets a benchmark also outside the territorial boundaries of the United States. The social and political relevance of the issue at stake, as well as the tentative guidelines offered by the USSC undoubtedly make this decision of relevance for all Contracting States to the 1980 Hague Convention, included those that are EU Member States. This paper will now focus on investigating the possible impact that this decision may have within the EU, where the legal framework is different.
While all EU Member States are Contracting States to the 1980 Hague Convention in their own right, the Convention is in some respects applied differently in relations among Member States for two reasons.
First, as is well-known, the 1980 Hague Convention was partially modified by the Brussels II-ter Regulation,34 which supplements and ‘complements’35 Convention provisions. A partial derogation from the Convention is allowed under Article 36 of the Convention, which permits two or more Contracting States to agree between themselves ‘to derogate from any provisions’ of the Convention ‘in order to limit the restrictions’ to which the return of the child may be subject under the Convention provisions.36 EU Member States have accomplished this by introducing common ‘special’ rules for intra-EU abduction cases.
While it is not possible to comment here on the Brussels II-bis and II-ter rules, it may be useful to recall that the general idea underlying the special rules enshrined in the Brussels II system is to reinforce the return obligation,37 something that has been claimed to protect the interests of States more than those of children (Freeman and Taylor 2022). This is accomplished in several ways. First, a court cannot refuse return unless the person seeking the return of the child has been given the opportunity to be heard (Article 27(1) Brussels II-ter). A genuine opportunity to be heard must be offered to the child as well (Article 26 Brussels II-ter), and this will have an impact, among other things, on the application of Article 13(2) of the 1980 Hague Convention. Secondly, the use of the grave risk exception under Article 13(1)(b) is limited. Even when the court is convinced that this defence is available, nonetheless, it ‘shall not refuse to return the child’ if it is satisfied that adequate arrangements have been made to secure the protection of the child upon his or her return (Article 27(2)). Such adequate arrangements may be proposed by the party seeking the return of the child, or by the court on its own initiative. Thirdly, even when the court in the State of refuge refuses the return, the court in the State of habitual residence may make use of the special mechanism under Article 29, empowering that court to adopt a decision on custody entailing the return of the child. Any such decision will override the decision on non-return in the State of refuge and must be enforced, subject only to limited exceptions.
The following analysis will focus on the role and special responsibility borne by the court in the State of refuge, as established by Article 27(3) and (5). These two provisions direct the State of refuge to ensure that adequate arrangements are taken in the State of habitual residence or to issue provisional protective measures enforceable in that State in order to safeguard the return of the child. Such measures will accompany the return of the child to his or her place of habitual residence and will produce effects and be enforced there.
Before turning to the legal analysis of the technical rules, however, it is important to note another key difference between the Regulation and the Convention. In contrast with the Convention, the Regulation sets out rules for both return proceedings and custody proceedings. Based on this structure, the Regulation creates a special relationship between the court of the State of habitual residence and the one in the State of refuge. Courts in different EU Member States are united by reciprocal trust in the level of legal protection granted in other Member States as well as the common effort to ensure the best solution for the child’s welfare. It is important to emphasise that EU Member States share common values, a similar level of protection for fundamental rights and comparable standards in the administration of justice. They also share and agree on common policies that have an impact on the application of the Convention, such as, for example, mainstreaming policies for the protection of children and protection against domestic violence. This circumstance qualifies the approach of EU Member States when dealing with intra-EU abduction cases. It is probably easier to order the return of a child if the sending court trusts the ability of the receiving court to assess the best interest of the child against standards similar to those that the sending court applies in its own forum.
Some twenty years have not passed since the EU became involved in combating violence against women, and only the most relevant instruments can be noted here. Following a request made by the European Parliament in 2010,38 the Commission proposed the adoption of Directive 2011/99/EU on the European protection order (EPPO)39 as well as Regulation No 606/2013 on mutual recognition of protection measures in civil matters.40 While both instruments are more general in scope, they were clearly aimed at providing protection to women who are exposed to intimate/domestic violence. Subsequently, as part of the Gender Equality Strategy 2020–2025,41 measures were announced to prevent forms of violence against women and domestic violence, protecting victims and prosecuting offenders. In 2022, the EU Commission launched a proposal for an EU directive on combating violence against women and domestic violence,42 which is currently passing through approval procedures. Finally, the Istanbul Convention on Preventing and Combating Violence Against Women and Domestic Violence43, which was signed by the EU as early as 2017, was finally approved by the Council on 1 June 2023 and will enter into force for the EU on 1 October 2023.44 The topic clearly remains at the top of the EU agenda.
On a different level, legal studies and research projects dealing with domestic violence in general, or more specifically, with the interaction between domestic violence and abduction,45 have in recent years contributed to increasing awareness and exploring more efficient solutions. Without repeating what has already been discussed elsewhere, one might summarise the following common arguments:
  • Domestic violence is being alleged with increasing frequency as a reason for leaving the State of common residence and relocating to a different Member State. When the woman is the primary caregiver, she will usually take the child/children with her, thus committing child abduction.
  • Providing convincing evidence of domestic violence, including its more subtle forms, such as financial abuse or coercive control, may prove to be extremely difficult, especially within the ambit of Hague return proceedings, as they involve summary proceedings with limited scope for investigation.
  • Even when domestic violence is committed against the mother and the child has not been physically harmed, the child is always an ‘indirect’ victim, having been exposed to the effects of domestic violence against the mother. This violence will undermine the psychological health and emotional well-being of the mother and thus her ability to parent and care for the child.
  • Although the 1980 Hague Convention protects (only) the child from the risk of harm upon return, it is now recognised that witnessing domestic abuse causes substantial harm that can have long-term effects on the welfare and development of children, leading to post-traumatic stress disorder and behavioural issues. This amounts to a psychological harm or an intolerable situation for the child, which is covered by Article 13(1)(b).46
Against this backdrop, the Brussels II-ter Regulation seeks to strike a balance between taking account of the need to protect the child and the mother, and the opposite need to comply with the Convention’s obligation to return the abducted child, both building on and innovating beyond the previous experience of the Brussels II-bis Regulation.

8. The Responsibility of the State of Refuge under Brussels II-ter: From ‘Adequate Arrangements’ to ‘Provisional and Protective Measures’

As with the Convention, the burden of the legal obligation and responsibility to decide whether or not to return the child lies with the State to which the child is removed. However, in contrast with the Convention, the Regulation requires that, when complying with this obligation, the court shall always ensure that the return of the child is ‘safe’. The concept of ‘safe return’ must be construed differently from that of the best interests of the child. Assessing the best interests of the child over the long run is not the primary concern of the courts of the State of refuge, and this task should be left to the courts of the State of habitual residence.
In order to do so, the Regulation equips the courts in the State of refuge with two additional tools, which are not expressly provided under the Convention: assessing adequate arrangements that could potentially be put in place in the State of habitual residence under Article 27(3), and directly adopting provisional protective measures under Article 15, which, according to Article 27(5), will be recognised and enforced exceptionally in the State of habitual residence.
Article 27(3) Brussels II-ter is modelled on the previous Article 11(4) Brussels II-bis and provides that, where a court considers refusing to return a child solely on the basis of Article 13(1)(b) of the Hague Convention, it cannot do so:
‘if the party seeking the return of the child satisfies the court by providing sufficient evidence, or the court is otherwise satisfied, that adequate arrangements have been made to secure the protection of the child after his or her return.’
Recital (45) clarifies what qualify as ‘adequate arrangements’ for the purposes of the rule. These include a court order from the State of the child’s habitual residence prohibiting the applicant from coming close to the child; a provisional and/or protective measure allowing the child to stay with the abducting parent who is the primary carer until a decision has been made on the merits in relation to custody; the indication of available medical facilities for a child in need of medical treatment.47 Other examples are mentioned in the 2022 Practice Guide (Practice Guide 2022, p. 125 at, such as secure accommodation for the parent and the child, the termination of criminal proceedings against the abducting parent, and covering the abducting parent’s living costs. These measures must be properly put in place in the given case, and not simply generally available under the law of the State of habitual residence (Practice Guide 2022, p. 125 at, further clarifying that judicial measures only need to be enforceable but not necessarily final).
While the content and breadth of such measures may depend on the circumstances of each individual case, a typical feature of all ‘adequate arrangements’ is that they must be adopted by an authority—a court, a child welfare authority, a social service—in the State of habitual residence of the child. In practice, these measures are mostly sought by the party seeking the return of the child or are offered by the Central Authority of the State of habitual residence. They may also be suggested by the court of the State of refuge on its own motion, but even in this case, the court has no power to craft or order them and must rely on arrangements made by and in the State of habitual residence.48 On the other hand, such measures do not raise any PIL issue with regard to their recognition or enforcement, as they are ordered by an authority of the State where they will be enforced.
On the contrary, it falls within the competence and duties of the court in the State of refuge to assess whether the proposed arrangements are ‘adequate’, i.e., effective for the purpose of securing a safe return.49 Interestingly, the State of refuge is not under any obligation to establish adequate arrangements, which by contrast are simply one of the various legal tools the court can—but not necessarily needs to—rely on in order to comply with the State’s international obligations (for a different view, see Chalas 2023). The point is not addressed in the Regulation itself and has never been analysed by the CJEU. However, the 2022 Practice Guide clearly affirms that ‘Adequate arrangements may be considered by the court of first instance or by the court of the higher instance in the Member State of refuge’ (Practice Guide 2022, p. 125 at In sum, in many respects, including this last relevant point, ‘adequate arrangements’ may be said to correspond to the ‘ameliorative measures’ within US legal practice noted above.
Article 27(5), by contrast, will have a much more powerful impact. This provision should be regarded as one of the most significant innovations introduced by the Brussels II-ter Regulation in the area of international child abduction. This new rule vests the courts in the State of refuge with an exceptional power to take protective measures that have extraterritorial effects and are enforceable in all EU Member States.50 The gap in protection that is inherent to the Hague Convention,51 and which also existed under the previous Brussels II-bis Regulation, which did not allow such protective measures to circulate, has now been filled by the new provisions of the Brussels II-ter Regulation (Wilderspin 2022).
The rule is often described as a further attempt to deter the courts of the State of refuge from refusing to return the child because of a grave risk of harm (Practice Guide 2022, p. 124 at 4.3.5). This is, however, only one side of the story. In the present author’s view, it is not only the obligation to return the child notwithstanding the existence of a risk of harm that is reinforced. Far more than this, it is the obligation to ensure that the child’s return is actually safe that is spelled out. The provision imposes a positive obligation on the court of the State of refuge to take steps to protect the child (Honorati 2022, p. 159). In other words, rather than expecting the State of habitual residence to take adequate arrangements, or waiting for the parent left behind to provide sufficient evidence that such measures are in place, the court must assess the child’s needs and, where appropriate, take any necessary protective measures itself. This also implies that when such measures are not available, either because they are not possible in concreto or because they are not effective as they would not be capable of sufficiently reducing the risk of harm, then the court has no alternative but to refuse the return of the child.
Content-wise, provisional measures will very much resemble adequate arrangements (Wilderspin 2022, p. 185). Recital (46) clarifies that such measures can, for example, provide that, once returned, the child will continue to reside with the primary caregiver, or specify how contact with the parent left behind should take place after the child’s return. Other examples that have been given include measures akin to anti-molestation/anti-harassment orders (for example, ‘not to use violence or threats towards the mother, nor to instruct anybody else to do so, or not to communicate with the mother directly), orders related to the occupancy of the family home (for example, to vacate the family home and make it available for sole occupancy by the mother and the child), orders related to financial support (for example, to pay for the return tickets for the mother and the child or to provide financial support/maintenance to the mother and the child upon their return), and orders related to residence or access to the child (for example, not to seek to separate the mother from the child or not to seek contact with the child unless awarded by the court or agreed) (Momoh 2022, p. 77). Reference may also be usefully made to the Guide to Good Practice on Article 13(1)(b) (HCCH 2020, para 43), according to which protective measures can cover a broad range of existing services, assistance, and support, including access to legal services, financial assistance, housing assistance, health services, shelters, and other forms of assistance or support to victims of domestic violence. It should be noted that, in a similar manner to the caution called upon by the USSC in the Golan decision, recital (46) recommends that protective measures should not ‘undermine the delimitation of jurisdiction between the court seised with the return proceedings under the Hague Convention and the court having jurisdiction on the substance of parental responsibility under this Regulation.’52
All in all, in contrast to what occurs in a purely Conventional situation, the new legal framework established by the Brussels II-ter Regulation not only allows but, indeed, requires the court of the State of refuge to take positive action—and responsibility—in order to protect the child from any kind of harm that he or she may suffer upon return.
Notwithstanding this new approach, however, there still is an important gap in terms of protection. This concerns the person who is exposed to domestic violence. The clear wording of Article 13(1)(b) leaves no doubt that the risk of physical and psychological prejudice must apply to the child, not to the mother. Although psychological studies show beyond any reasonable doubt the devastating impact of domestic violence on children, even if they have not witnessed the violence (among many: Lindhorst and Edleson 2012; Katz 2022; POAM Best Practice Guide 2022),53 according to a literal interpretation of Article 27(5), doubts may arise as to whether this provision allows for the adoption of protective measures with regard to a situation involving primarily the mother. It is regrettable that, despite calls from particularly attentive scholars (Trimmings 2013, p. 154), EU lawmakers have not taken the opportunity to clarify this matter. A reference to domestic violence could, at least, have been included in a recital. Instead, although it does provide examples of possible protection measures, recital (46) does not refer, even implicitly, to situations that could involve the mother. As has been note elsewhere:
‘the mother is left alone to face a terrible dilemma: either return with the child and go back to the situation of violence she had escaped from, or stay safe and protected, but abandon her child’.
While legal scholars who have tackled the issue have supported the view that the rule should be construed as also envisaging measures for the protection of the mother (Wilderspin 2022, p. 186; Honorati and Ricciardi 2022, p. 252; and also, although with a different approach Requejo Isidro 2006; Ripley 2008), the current lack of a clear indication and the subsequent flaw in the instrument will only result in the confirmation of the existing tendency before the courts. In most cases, in fact, Hague courts tend to overlook the effects of domestic and intimate violence on mothers and rarely recognise the psychological harm suffered by the child(ren). The sparse research that has investigated how courts use adequate arrangements in practice show that these have been used only in a limited number of cases, especially when compared to the high proportion of cases involving allegations of domestic violence (Trimmings 2013, p. 155 et seq.; Honorati 2020, p. 817).54
This situation should, instead, be handled and considered by the law. It is submitted here that, under the current legal framework, in most cases where there is an allegation of domestic violence, the State of refuge should always consider—and possibly adopt—some kind of protective measure in order to accompany the child (and the mother) back home safely. This should become part of a settled routine before the courts. Where the court is not fully convinced of the existence of a grave risk of harm but still cannot exclude that the situation may entail some kind of additional risk, the possibility of so-called ‘soft landing’ measures should be considered. What is important is that the Hague court—which is often a specialised court that deals, mainly or exclusively, with abduction cases—refrains from dismissing an allegation of domestic violence on the mere assumption that such a court should focus on the child only.

9. How Protective Are ‘Protective Measures’?

Before proceeding to the conclusions, an additional disclaimer must be made. The consideration of ameliorative measures or the ordering of protective measures will not always be sufficient to resolve the case. Sometimes they will not be enough.
This issue engages two different levels of analysis. The first one is purely legal. It considers how protective measures can be legally binding and effective—or better: enforceable—in a State different from the State that adopted them. As seen, this issue is now addressed and resolved within the EU by the new Article 27(5) Brussels II-ter Regulation. Outside the EU, courts will need to use some creativity as the 1980 Hague Convention does not deal with the recognition of decisions and, hence, does not provide for a solution. The more obvious solution is to use other international treaties on the recognition of decisions that may be in force between the relevant States. If, for example, both States are Contracting States to the 1996 Hague Convention, Article 11 may be invoked. In other cases, courts may either use mirror-orders—i.e., orders with the same content that are issued in both the State of refuge and the State of habitual residence; or safe harbour orders—i.e., orders issued by the State of refuge or by the State of habitual residence stipulating certain conditions for a safer and less disruptive return of the child. Courts may thus call upon the parties (or the Central Authority) to ask the courts in the State of the child’s habitual residence to adopt adequate measures. This is what happened in the Golan case, where both parties applied to the Court of Milan seeking a package of measures to be applied upon the return of B.A.S.
Besides the legal issues, however, there is a second level of more substantial concern, which in the opinion of the present author has not been sufficiently considered by legal scholars. This relates to the fact that in some cases, with regard to specific situations involving serious domestic violence, there is no way to protect the victim because the tortfeasor is incapable of controlling his behaviour and complying with a binding court order. This argument was repeated in Narkis Golan’s defence and in some of the Amici Curiae Briefs (such as Brief for Amici Curiae 2021). In particular, it has been stressed that
‘Domestic abuse is sometimes mistakenly understood as a series of discrete violent acts, when in fact it is most often an insidious pattern of physical and psychological abuse marked by an ever-present exploitation of control. […]
Perpetrators of domestic abuse use a combination of tactics to maintain and gain power and control over their target, including but not limited to physical, sexual, psychological, emotional, economic, and immigration-related abuse. Using a combination of these modes of abuse, perpetrators gradually begin to exert an insidious but powerful kind of manipulative control over their victims, known as “coercive control”. […]
Efforts to craft ameliorative measures are based on the often-erroneous assumption that the abuser, […] will reform and start to live consistent with a set of conditions wholly out of step with the abuser’s past conduct. In reality, serious and persistent abusers generally do not abandon their abusive conduct, especially when there is no criminal penalty imposed or close monitoring of their behavior.’
This is not the appropriate place to define the terms and standards that determine whether and when a violent and abusive man is capable of changing and complying with a court order. It is also acknowledged that this may be a difficult task for a legal scholar. It cannot be denied, however, that there are cases in which there is no appropriate legal protection against an abuser seeking to ‘punish the victim’s efforts to escape and to re-establish control through even stronger means’ (Brief for Amici Curiae 2021, p. 10). It must, instead, be recognised that, in such cases, the only possible defence for the victim is to escape as far as possible. It is certainly not an easy task for a court to separate these particular cases (which are potentially limited in number) from the majority of cases in which a protected return would be the appropriate solution. There is no doubt that a greater awareness not only of the effects of domestic violence on women and children but also of the behavioural dynamics of abusers would be of great help to courts in this difficult task.

10. Conclusions—Advancing the Protection of the 1980 Hague Convention: A Lost Opportunity for the USSC, and a Bad Example for the EU

The issue addressed in this article, i.e., how to deal with abduction cases that are motivated by domestic violence, seems to divide legal scholars into two mainstream camps. On one side, there are ‘feminist’ lawyers, sometimes also criminal lawyers mostly active in advocating domestic violence cases, who at times appear to overlook or undervalue the merits and structure of one of the most successful conventions, which is in force among an exceptionally high number of States. On the other side, we have ‘internationalist’ scholars or Hague Convention ‘purists’ who appreciate the overall structure and consider domestic violence cases as one single piece in the broader puzzle, which must not undermine the solidity of the general framework.
The time has now come to bridge the gap between the two extremes and the needs associated with each of them. Each and any reasonably supported allegation of domestic violence must (and not only should) always be taken into consideration by the courts. It is today no longer acceptable that a defence, which is based on alleged domestic violence and shows some piece of evidence, is overlooked or not seriously taken into consideration by the courts, even where this occurs within summary return proceedings and even where (or rather especially where) the evidence provided is poor. Courts are under a duty to guarantee the safety of the child—not only physical safety but also psychological safety as resulting from the emotional balance of the primary caregiver—and this outcome should not be made dependent on the ability of the abducting mother’s counsel to argue or to provide sufficient evidence of the case.
Against this backdrop, it is submitted here that this gap could be bridged by relying on ameliorative/protective measures. In most cases (though by no means not in all cases), expeditious, substantively well-defined, and enforceable protection measures will offer the best guarantee that domestic violence cases are not neglected, showing consideration for the risk alleged by the abducting/escaping mother, while at the same time moving within the scope of the 1980 Hague Convention. Of course, a key consideration is that such measures must be effective and enforceable in the State of habitual residence. As has been noted:
The need for cross-border protective measures in return orders has become an essential part of the fabric of 1980 Convention proceedings, and ensuring that we have the right tools for recognition and enforcement is key’.
While under the new Regulation this is already the case within the EU, something more needs to be done for other Contracting States. The solution potentially envisaged by the USSC in the Golan case is legally sound and should be endorsed—as no Convention provisions formally require any ameliorative or protective measures. Nonetheless, the wrong message may have been given overall, namely that ameliorative measures are not necessary/useful and, at the same time, that domestic violence is not a priority issue for the courts. It is submitted here that the USSC lost the opportunity to send out a clear message that domestic violence is a plague that needs to be fought at all levels and also to provide clearer directions on how to do so within the framework of the Hague Convention. Making continuous references to a literal interpretation of international agreements may have dangerous secondary effects, as this may promote the interpretation that violence committed against the mother is irrelevant. Instead, this would have been a good opportunity to show how to keep this fundamental convention up to date with a changing society and to ‘nurture’ (Freeman and Taylor 2023) it accordingly so as to stand up for the protection of women.


This research received no external funding.

Conflicts of Interest

The author declares no conflict of interest.


  1. Barnett, Adrienne. 2023. The Hague Abduction Convention: What Needs to Change? Available online: (accessed on 1 June 2023).
  2. Beaumont, Paul, Katarina Trimmings, Lara Walker, and Jane Holliday. 2015. Child Abduction: Recent Jurisprudence of the European Court of Human Rights. International and Comparative Law Quarterly 64: 39–63. [Google Scholar] [CrossRef]
  3. Bergamini, Chiara. 2023. The proposal for a EU directive on combating violence against women and domestic violence. In Migrant Women and Gender-Based Violence in the International and European Legal Framework. Edited by Angela Di Stasi, Raffaele Cadin, Anna Iermano and Valentina Zambrano. Napoli: Editoriale Scientifica, p. 487 ss. [Google Scholar]
  4. Brief for Amici Curiae. 2021. Brief for Amici Curiae Individuals and Organizations Advocating for victims of Domestic Violence in Support of Petitioner. Available online: (accessed on 1 June 2023).
  5. Brief for Amici Curiae. 2022. Brief of Amici Curiae, Professors of Law Linda J. Silberman, Robert G. Spector, and Louise Ellen Teitz, in Support of Respondent. Available online: (accessed on 1 June 2023).
  6. Chalas, Christelle. 2023. Les measures de protection de l’enfant illicitement déplacé et le risque grave de danger: Comparaison de l’office des juges américains et européens. Revue Critique de Droit international Privé 1: 37–84. [Google Scholar]
  7. Fidler, Nicole. 2023. In Memory of Narkis Golan. Family Law Quarterly 56: 271–73. [Google Scholar]
  8. Freeman, Marilyn, and Nicola Taylor. 2022. Domestic violence as an aspect of 1980 Hague Child Abduction Convention return proceedings. In Domestic Violence and Parental Child Abduction: The Protection of Abducting Mothers in Return Proceedings. Edited by Katarina Trimmings, Anatol Dutta, Costanza Honorati and Mirela Župan. Cambridge: Intersentia Ltd., pp. 39–65. [Google Scholar]
  9. Freeman, Marilyn, and Nicola Taylor. 2023. Contemporary Nurturing of the 1980 Hague Convention. Laws 12: 65. [Google Scholar]
  10. HCCH. 2017. Draft Guide to Good Practice under the Hague Convention of 25 October 1980 on the Civil Aspects of International Child Abduction, Part VI, Article 13(1)(b). Prel. Doc. No 3 of June 2017. Available online: (accessed on 1 June 2023).
  11. HCCH. 2020. Guide to Good Practice under the Hague Convention of 25 October 1980 on the Civil Aspects of International Child Abduction, Part VI, Article 13(1)(b). The Hague: HCCH Permanent Bureau. [Google Scholar]
  12. Honorati, Costanza. 2020. Il ritorno del minore sottratto e il rischio grave di pregiudizio ai sensi dell’art. 13 par. 1 lett. b della Convenzione dell’Aja del 1980. Rivista di Diritto Internazionale Privato e Processuale 4: 796–842. [Google Scholar]
  13. Honorati, Costanza. 2022. Jurisdiction to take protective measures in the State of refuge in the context of child abduction. In Domestic Violence and Parental Child Abduction: The Protection of Abducting Mothers in Return Proceedings. Edited by Katarina Trimmings, Anatol Dutta, Costanza Honorati and Mirela Župan. Cambridge: Intersentia Ltd., pp. 139–60. [Google Scholar]
  14. Honorati, Costanza, and Giovanna Ricciardi. 2022. Violenza domestica e protezione crossborder. Rivista di Diritto Internazionale Privato e Processuale 2: 225–56. [Google Scholar]
  15. Katz, Emma. 2022. Coercive Control in Children’s and Mothers’ Lives. Oxford: Oxford University Press. [Google Scholar]
  16. Lindhorst, Taryn, and Jeffrey Edleson. 2012. Battered Women, Their Children, and International Law: The Unintended Consequences of the Hague Child Abduction Convention. Boston: Northeastern University Press. [Google Scholar]
  17. Mc Eleavy, Peter. 2015. The European Court of Human Rights and the Hague Child Abduction Convention: Prioritising Return or Reflection? Netherlands International Law Review 62: 365–405. [Google Scholar] [CrossRef]
  18. Momoh, Onyoja. 2022. The need for cross-border protective measures in return proceedings. In Domestic Violence and Parental Child Abduction: The Protection of Abducting Mothers in Return Proceedings. Edited by Katarina Trimmings, Anatol Dutta, Costanza Honorati and Mirela Župan. Cambridge: Intersentia Ltd. [Google Scholar]
  19. Pahrand, Mishal. 2017. Not Without my Children: The Need for the Modification of International Child Abduction Laws. Family Court Review 55: 139–51. [Google Scholar] [CrossRef]
  20. POAM Best Practice Guide. 2022. Best Practice Guide—Protection of Abducting Mothers in Return Proceedings, reprinted. In Domestic Violence and Parental Child Abduction: The Protection of Abducting Mothers in Return Proceedings. Edited by Katarina Trimmings, Anatol Dutta, Costanza Honorati and Mirela Župan. Cambridge: Intersentia Ltd. Available online: (accessed on 1 June 2023).
  21. Practice Guide. 2022. Practice Guide for the Application of the Brussels IIb Regulation. Luxembourg: EC Commission. [Google Scholar]
  22. Requejo Isidro, Marta. 2006. Secuestro de menores y violencia de género en la Union Europea. Anuario Español de Derecho Internacional Privado 6: 190–92. [Google Scholar]
  23. Ripley, Peter. 2008. A Defence of the Established Approach to the Grave Risk Exception in the Hague Abduction Convention. Journal of Private International Law 4: 443–77. [Google Scholar] [CrossRef]
  24. Trimmings, Katarina. 2013. Child Abduction within the European Union. Oxford: Hart Publishing. [Google Scholar]
  25. Trimmings, Katarina, Anatol Dutta, Costanza Honorati, and Mirela Župan, eds. 2022. Domestic Violence and Parental Child Abduction: The Protection of Abducting Mothers in Return Proceedings. Cambridge: Intersentia Ltd. [Google Scholar]
  26. Trimmings, Katarina, and Onyoja Momoh. 2021. Intersection between domestic violence and international parental child abduction: Protection of abducting mothers in return proceedings. International Journal of Law Policy and the Family 35: 1–19. [Google Scholar]
  27. Trimmings, Katarina, Konstantina Kalaitsoglou, Pinsent Masons, and Rima Hussein. 2023. Protection of Abducting Mothers in Return Proceedings under the 1980 Hague Convention on the Civil Aspects of International Child Abduction. Policy Brief, July. Available online: (accessed on 1 June 2023).
  28. Weiner, Merle. 2021. You Can and You Should: How Judges Can Apply the Hague Abduction Convention to Protect Victims of Domestic Violence. UCLA Women’s Law Journal 28: 223–332. [Google Scholar] [CrossRef]
  29. Wilderspin, Michael. 2022. The Potential Role of Regulation 606/2013 in Protecting the Abducting Parent in Return Proceedings. In Domestic Violence and Parental Child Abduction: The Protection of Abducting Mothers in Return Proceedings. Edited by Katarina Trimmings, Anatol Dutta, Costanza Honorati and Mirela Župan. Cambridge: Intersentia Ltd. [Google Scholar]
Golan v Saada, 596 U.S. (2022).
See below at Section 2.1.3.
The couple married in Tel Aviv in a religious ceremony, but the marriage was never registered in any country. As a result, Mrs. Golan was unable to work legally in Italy.
All quoted passages are from USDC (E.D.N.Y.) No. 18-5292 (22 March 2019).
USDC (E.D.N.Y.) No. 18-5292 (22 March 2019), p. 44a note 3, explaining how Mr. Saada’s grandparents, parents, brother, and sister were living in Milan on different floors of the same building.
Further details at USDC (E.D.N.Y.) No. 18-5292 (22 March 2019), pp. 56a–59a.
The argument raised by the mother that B.A.S. was habitually a resident in the US, as this was the shared intention of the parents, was quickly rejected.
See Blondin v Dubois case, which is discussed below.
These measures included: a $30,000 allowance for Mrs. Golan; staying away from her until the Italian courts had resolved the matter; starting cognitive behavioural therapy; discontinuing any criminal or civil proceedings that could be pursued in Italy in relation to abduction. In addition to this, Mr. Saada was also required to submit the full record of the New York proceedings to the Italian court deciding on the custody proceedings and to assist Mrs. Golan in obtaining legal status and working papers in Italy.
USDC (E.D.N.Y.), No. 18-5292 (19 March 2019).
USCA, 2nd Circuit, No. 19-820 (19 July 2019).
USDC (E.D.N.Y.), No. 18-5292 (5 May 2020).
USCA, 2nd Circuit, No. 20-1544 (28 October 2020).
All documents and Amicus Curiae Briefs can be found at the USSC docket accessed on 1 June 2023.
See, in particular, the following cases: Danaipour v McLarey, 386 F.3d 289 (1st Circuit, 2004); Acosta v Acosta 752 F. 3d 868 (8th Circuit, 2013); Baran v Baran, 526 F. 3d 1340 (11th Circuit, 2008); for details on decisions, see Petition for a Writ of Certiorari, Golan v Saada, January 2021, at 11–12.
Blondin v Dubois (I), 189 F.3s 240 (2d Cir. 1999).
Idem at 249–250.
Blondin v Dubois (II), 238 F.3d 153 (2d Cir. 2001) at 163 n. 11.
For example, In re Adam, 437 F 3d 381(3d Circuit, 2006); Gaudin v Remis, 415 F.3d 1028 (9th Circuit, 2005). See again Petition for a Writ of Certiorari, at 14. The Petition also points to further inconsistencies in case law of state courts at 17–18.
The principle is also rooted in Article 31 of the 1969 Vienna Convention on the Law of Treaties, to which, however, the USA is not a Contracting Party. Article 31, after setting the general rule that a treaty ‘shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose’, first clarifies what is the context in paragraph (2). Paragraph (3) further after adds that together with the context shall be taken into account: ‘(a) any subsequent agreement between the parties regarding the interpretation of the treaty or the application of its provisions; (b) any subsequent practice in the application of the treaty, which establishes the agreement of the parties regarding its interpretation; (c) any relevant rules of international law applicable in the relations between the parties’. The rule may be of relevance to other States which are bound by the Vienna Convention.
Golan v Saada, p. 10, footnote 7.
Golan v Saada, p. 9.
Golan v Saada, p. 10.
Golan v Saada, p. 10–11 (emphasis added).
Golan v Saada, p. 11.
See footnote 25.
Golan v Saada, p. 12.
Golan v Saada, p. 13.
See in particular (HCCH 2020) Section 1 on how to assess grave risk (p. 31, para 38 seq) and Section 2 providing examples of grave risk, including relevant cases of domestic violence, at para 55–76.
Again, the USSC refers for guidance to internal practice. See the letter of 10 August 1995 sent by the U. S. Dept. of State to the United Kingdom. See Golan, p. 13, note 9.
The US has only signed the 1996 Hague Convention on Jurisdiction, Applicable Law, Recognition, Enforcement and Co-operation in Respect of Parental Responsibility and Measures for the Protection of Children. The situation may be different for jurisdictions, such as the UK or Australia, which have ratified the Convention. For the list of Contracting States, see accessed on 1 June 2023. The situation is again different with regard to EU Member States applying the Brussels II-ter Regulation. See below at Section 0.
See the Hague Mothers project ( accessed on 1 June 2023). The project’s website states: ‘Our overarching aim is to end the injustices created by The Hague Convention, specifically for mothers and children who are fleeing abusive relationships’. The aim of ‘Phase 4’ is to ‘Amend the Convention and/or the regulations to protect mothers and children who have fled violence or abuse’. National regulations have already been amended in order to take domestic violence into account in Australia, Switzerland, and Japan.
Letter of the Children’s Law Center addressed to Judge Donnelly, 3 November 2022.
Council Regulation (EU) 2019/1111 of 25 June 2019 on jurisdiction, the recognition and enforcement of decisions in matrimonial matters and the matters of parental responsibility, and on international child abduction (recast)—known as the Brussels II-ter Regulation.
Article 22 Brussels II-ter Regulation.
Article 36 of the 1980 Hague Convention provides that: ‘Nothing in this Convention shall prevent two or more Contracting States, in order to limit the restrictions to which the return of the child may be subject, from agreeing among themselves to derogate from any provisions of this Convention which may imply such a restriction.’
See generally the system of Regulation No 2201/2003 (Brussels II-bis) and its recast Regulation No 2019/1111 (Brussels II-ter), insofar as the rules of the two instruments are inspired by the same rationale. The following paragraph will focus on the rules currently in force under Brussels II-ter.
The European Parliament resolution of 26 November 2009 on the elimination of violence against women, [2010] OJ EC 285/53 of 21 October 2010; the European Parliament resolution of 5 April 2011 on priorities and outline of a new EU policy framework to fight violence against women, [2012] OJ EC 296/26 of 2 October 2012.
Directive 2011/99/EU of the European Parliament and of the Council of 13 December 2011 on the European protection order.
Regulation (EU) No 606/2013 of the European Parliament and of the Council of 12 June 2013 on mutual recognition of protection measures in civil matters.
COM(2020)152 final.
COM(2022) 105 final of 8 March 2022. The proposal focuses more on a criminal law approach, even though violations of human rights and forms of discrimination are also considered. On the proposal and for further references, see (Bergamini 2023).
The Convention is promoted by the Council of Europe and was opened for signature on 11 May 2011; it is currently in force today in 37 Contracting States. It has been signed by all EU Member States, and 21 of them have also ratified it in their own right. (see accessed on 1 June 2023). For the position of the EU see below.
Council Decision (EU) 2023/1075 and Council Decision (EU) 2023/1076 of 1 June 2023, on the conclusion, on behalf of the European Union, of the Council of Europe Convention on preventing and combating violence against women and domestic violence, respectively, with regard to institutions and public administration of the Union. The underlying rationale for the two decisions is that both the EU and Member States have competence in the fields covered by the Convention, and the EU can only adopt the Convention with regard to its own sphere of competence. Furthermore, a Code of Conduct was also adopted, setting out the internal arrangements for practical cooperation and cooperation between the EU and the Member States on various aspects of the implementation of the Convention (ST/8113/2023/INIT, in [2023] OJ C 194/03). The decision of the Council authorising the ratification is supposed to put some pressure on the six remaining EU Member States: Bulgaria, Czechia, Hungary, Latvia, Lithuania, and Slovakia.
See the POEM (Mapping the legislation and assessing the impact of Protection Orders in the European Member States) and the POAM (Protection of Abducting Mothers in Return Proceedings: Intersection between Domestic Violence and Parental Child Abduction) projects. The results of the former are accessible here: accessed on 1 June 2023, and those of the latter here: accessed on 1 June 2023 and in Trimmings et al. (2022)). See also POAM Best Practice Guide 2022.
As the UK Supreme Court held in Re E (Children) [2011] UKSC 27, ‘it is now recognised that violence and abuse between parents may constitute a grave risk to the children’. A child should thus not reasonably be expected to tolerate ‘exposure to the harmful effects of seeing and hearing the physical or psychological abuse of her own parent’ (para. 52 and 34).
Useful reference can also be made to HCCH 2020, paras 43 et seq.
This does not mean that the Court in the State of refuge is entirely passive. Indeed, it may play an active role in the establishment of such measures, and it can also act ex officio. However, there is no formal requirement to do so. The situation is pretty much similar to the one faced by the USSC in the Golan case.
Article 27(4) provides that, for the purposes of investigating the adequacy of any adequate arrangement, the court ‘may communicate with the competent authorities of the Member State where the child was habitually a resident before the wrongful removal or retention’ either directly or with the assistance of Central Authorities (see, however, recital (45), which advises that the court ‘should primarily rely upon the parties.’). The 2022 Practice Guide, p. 125 at acknowledges that ‘It may be difficult for the judge to establish what possible arrangements exist in the Member State of origin, if they have been de facto taken and whether they are adequate to deal with the circumstances that could develop after the return’.
This result does not stem clearly from the rule but is reached indirectly through a rather cumbersome referral to the definition of ‘decision’. According to Article 2(1)(b), for the purposes of recognition and enforcement, the notion of decision includes ‘provisional, including protective, measures ordered by a court, which by virtue of this Regulation has jurisdiction as to the substance of the matter or measures ordered in accordance with Article 27(5) in conjunction with Article 15’ (emphasis added). On the other hand, provisional measures taken pursuant to Article 15 by a court that has no jurisdiction as to the substance of the matter will not have any effect outside that State. It must be noted however, that provisional measures will be recognised and enforced in another Member State provided that the party against whom they are taken has been summoned to appear, or at least the decision containing the measure was served on that party prior to enforcement.
In the Golan case, for example, the Court of Appeals for the Second District initially vacated the District Court decision as ‘many of the undertakings the District Court imposed [on the petitioner Saada] are unenforceable because they need not—or cannot—be enforced until after B.A.S. is returned to Italy’, and they were, therefore, considered insufficient to protect the child (USCA, 19 July 2019).
A point also underlined by the 2022 Practice Guide at 4.3,5.2.
On the EU level, see the Opinion of the European Economic and Social Committee on Children as indirect victims of domestic violence, [2006] OJ C 325, pp. 60–64.
Reference should also be made to the Parent Survey conducted under the aegis of the POAM project from February to April 2021 and investigating cases of alleged domestic violence (at accessed on 1 June 2023). The survey found that, in 83% of the cases investigated, protective measures were not available, advised, or discussed. It also found that the mothers interviewed felt quite strongly about this professed injustice (Parent Survey Report para 22).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Honorati, C. Protecting Mothers against Domestic Violence in the Context of International Child Abduction: Between Golan v Saada and Brussels II-ter EU Regulation. Laws 2023, 12, 79.

AMA Style

Honorati C. Protecting Mothers against Domestic Violence in the Context of International Child Abduction: Between Golan v Saada and Brussels II-ter EU Regulation. Laws. 2023; 12(5):79.

Chicago/Turabian Style

Honorati, Costanza. 2023. "Protecting Mothers against Domestic Violence in the Context of International Child Abduction: Between Golan v Saada and Brussels II-ter EU Regulation" Laws 12, no. 5: 79.

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop