F. v. N., 2022 SCC 51

SUPREME COURT OF CANADA
Citation: F. v. N., 2022 SCC 51			Appeal Heard: April 12, 2022 Judgment Rendered: December 2, 2022 Docket: 39875
Between:   F. Appellant   and   N. Respondent   - and -   Attorney General of Ontario, Office of the Children’s Lawyer, Defence for Children International-Canada and Canadian Council of Muslim Women Interveners       Coram: Wagner C.J. and Moldaver, Karakatsanis, Côté, Brown, Rowe, Martin, Kasirer and Jamal JJ.
Reasons for Judgment: (paras. 1 to 137)	Kasirer J. (Wagner C.J. and Moldaver, Côté and Rowe JJ. concurring)
Dissenting Reasons: (paras. 138 to 196)	Jamal J. (Karakatsanis, Brown and Martin JJ. concurring)

 

 

Note: This document is subject to editorial revision before its reproduction in final form in the Canada Supreme Court Reports.

 

 

F.                                                                                                                      Appellant

v.

N.                                                                                                                   Respondent

and

Attorney General of Ontario,

Office of the Children’s Lawyer,

Defence for Children International-Canada and

Canadian Council of Muslim Women                                                       Interveners

Indexed as: F. v. N.

2022 SCC 51

File No.: 39875.

2022: April 12; 2022: December 2.

Present: Wagner C.J. and Moldaver, Karakatsanis, Côté, Brown, Rowe, Martin, Kasirer and Jamal JJ.

on appeal from the court of appeal for ontario

                    Family law — Custody — International child abduction — Jurisdiction to make parenting order — Serious harm to child — Best interests of child — Return order — Parties residing in United Arab Emirates with their two children — Mother taking children on trip to Ontario with father’s consent but refusing to return — Father seeking order from Ontario court for children’s return — Mother requesting that Ontario court exercise jurisdiction to make parenting order on merits — Ontario court declining jurisdiction on basis that it was not satisfied that children would suffer serious harm if removed from Ontario and ordering that children be returned to United Arab Emirates — Whether Ontario court erred in declining jurisdiction and ordering children’s return — Children’s Law Reform Act, R.S.O. 1990, c. C‑12, ss. 23, 40.

                    The father and the mother have two children, born in 2016 and 2019. The father is a citizen of Pakistan and has lived in Dubai, in the United Arab Emirates (“UAE”), since 2008. The mother, a citizen of Pakistan and of Canada, moved with her family from Pakistan to Ontario in 2005, and then to Dubai in 2012 when she married the father. The mother does not have an independent residency status in Dubai and has been sponsored by the father throughout their marriage. She is the children’s primary caregiver.

                    In June 2020, the mother travelled to Ontario with the children, leaving Dubai with return air tickets, ostensibly for a trip to visit her family. The father agreed to that trip; however, a few weeks later, the mother informed the father that she intended to stay in Ontario with the children and not return to Dubai. The father initiated proceedings in Ontario, seeking an order under s. 40 of the Children’s Law Reform Act (“CLRA”) for the return of the children to Dubai. The mother, relying on s. 23 of the CLRA, said that the Ontario court should exercise its jurisdiction to decide custody and access as the children would suffer serious harm if they were returned to Dubai. She claimed that it was in the best interests of the children to remain in Ontario with her. The father replied that the Ontario court should not take up jurisdiction to make a parenting order and that it was in the best interests of the children that all issues of custody and access be decided in the UAE.

                    Prior to the hearing of the father’s application, he presented a settlement offer to the mother in which he undertook to ensure the mother’s independent residency in Dubai, in particular, by purchasing property for her in her name. He further undertook that the children would reside primarily with the mother and that major decisions regarding the children would be made jointly.

                    The Ontario court declined jurisdiction. The trial judge was not satisfied that the children would suffer serious harm if they were removed from Ontario. He declared that the mother had wrongfully retained the children in Ontario and concluded that it was in the best interests of the children to return to Dubai, with or without the mother. He offered all parties the opportunity to make further submissions on whether to include the father’s settlement proposal in his order but the mother did not avail herself of this opportunity.

                    A majority of the Court of Appeal confirmed the return order, but the dissenting judge was of the view that the trial judge had erred in his assessment of serious harm and that the Ontario court should have exercised jurisdiction.

                    Held (Karakatsanis, Brown, Martin and Jamal JJ. dissenting): The appeal should be dismissed.

                    Per Wagner C.J. and Moldaver, Côte, Rowe and Kasirer JJ.: The trial judge in the instant case committed no reviewable error and his decision that the serious harm threshold was not met is entitled to deference. The custody dispute, which is undecided here, should be resolved by the courts in the UAE, where the children have their closest connection.

                    The CLRA seeks to discourage child abductions and the wrongful removal and retention of children to Ontario. The statute is based on the premise that, following an abduction, the child’s best interests are usually aligned with their prompt return to the jurisdiction of their habitual residence. Therefore, where a child who is wrongfully removed to or retained in Ontario habitually resides in a country that is not a party to the Convention on the Civil Aspects of International Child Abduction (“Hague Convention”), the CLRA provides that, but for exceptional circumstances, courts will refrain from exercising jurisdiction and leave the merits to the foreign jurisdiction with which the child has a closer connection. One exception is set forth in s. 23 of the CLRA: a court can exercise jurisdiction to make a parenting order where a child is physically present in Ontario and, on a balance of probabilities, the court is convinced that the child would suffer serious harm if removed from the province. At the preliminary stage of deciding jurisdiction, it is not the role of the judge to conduct a broad-based best interests inquiry, as they would on the merits of a custody application. A broad‑based best interests analysis under s. 23 would ultimately undermine the purpose of the serious harm exception, that is, to ensure decisions on the merits are made by the appropriate authority in accordance with the best interests of the child.

                    The onus to prove that the child would suffer serious harm upon return rests on the abducting parent. The burden is demanding and it is not enough to conclude that the return would have a negative impact on the child. It is also not enough to identify a serious risk of harm: the court must be satisfied, on a balance of probabilities, that the harm itself would be serious in nature. Serious harm inquiries are child‑centered and the analysis is highly individualized. The child’s age, and where relevant, their special needs and vulnerabilities, may mitigate or aggravate the risk of harm. When conducting their s. 23 analysis, judges should consider both the likelihood and severity of the anticipated harm. The focus is on the particular circumstances of the child, rather than on a general assessment of the society to which they would be sent back. Given the discretionary, individualized and fact‑specific character of the serious harm analysis, a trial judge’s findings are owed deference. Appellate courts cannot set aside a trial judge’s conclusion on serious harm simply on the basis that they would have weighed the evidence differently.

                    One relevant question as to the scope of the s. 23 exception is whether separation of a child from their primary caregiver can pose a risk to the child’s psychological well‑being rising to the level of serious harm. Separating an infant from their primary caregiver is a circumstance that most certainly can cause psychological harm to the child. But such a separation, in and of itself and without regard to the individualized circumstances, will not always rise to the level of harm required under s. 23 of the CLRA. In order to deter and remedy child abductions effectively, courts should sometimes be prepared to order the return of the children despite a risk of separation from their primary caregiver. Deciding otherwise could allow abducting parents, in some situations, to rely on their status as primary caregivers to circumvent the due process for custody determination and remove the children from the authority of the courts that would normally have jurisdiction. This could ultimately risk making Ontario a haven for child abductions.

                    When considering risks of harm flowing from separation, courts should recognize that if a child is separated from their primary caregiver, but is nevertheless returned to their capable left‑behind parent and other known caregivers, in a safe and familiar environment, the high threshold of harm may not be met. Courts should also consider all barriers to the return of the primary caregiver. Nevertheless, a parent ought not to be able to create serious harm and then rely on it through their own refusal to return. Courts should carefully scrutinize refusals to return when there is no impediment to the parent re‑entering and remaining in the country of the child’s habitual residence. But a primary caregiver’s refusal to return will not always be taken to be unjustified. An abducting parent may have legitimate and reasonable reasons for not returning to the foreign country, such as significant obstacles to employment or risks to safety, including evidence showing that the left‑behind parent is responsible for child abuse or intimate partner violence to the primary caregiver.

                    Another relevant question as to the scope of s. 23 is whether inconsistencies between family law in the foreign jurisdiction and in Ontario should factor in a serious harm analysis. As long as the ultimate question of custody will be determined by the court that has jurisdiction to do so on the basis of the best interests of the child, inconsistencies between local and foreign legal regimes will usually not amount to serious harm. Nonetheless, there may be instances where foreign laws are so profoundly irreconcilable with Ontario law that remitting the matter to the foreign courts would constitute serious harm within the meaning of the CLRA.

                    When a court is satisfied that a child has been wrongfully removed to or is wrongfully retained in Ontario, a return order presented by the left‑behind parent is governed by s. 40 of the CLRA. Judges should consider the best interests of the child in exercising their s. 40 powers. The return order procedure starts from the premise that the best interests of the child are aligned with their prompt return to their habitual place of residence so as to minimize the harmful effects of child abduction. If the evidence is insufficient to establish that Ontario courts should assume jurisdiction, judges should not use their residual s. 40 powers to postpone indefinitely the child’s return to the jurisdiction best positioned to decide the case on the merits.

                    Incorporating undertakings from the parties within a s. 40 return order may effectively facilitate a child’s return by providing an answer to an anticipated risk of harm. Even without a risk of serious harm within the meaning of s. 23, undertakings may be in the child’s best interests in that they effectively mitigate less consequential or short‑term distress. Problems associated with the enforceability of undertakings before foreign courts are well known. What is required is that the judge who hears the parties is satisfied that the undertakings given are adequate. This assessment is discretionary and must be made in light of the parties’ particular circumstances.

                    In the instant case, the trial judge did not commit a palpable and overriding error when he concluded that the children would not suffer serious harm if they were returned to Dubai. He understood that the separation of children from their primary caregiver typically gives rise to emotional distress for very young children. But he found, on the basis of the evidence, that this distress did not rise to the higher level of serious harm.

                    The allegations of serious harm faced by the children in the instant case relate, in part, to the risk that the mother, as primary caregiver, will be separated from them should they be ordered home to Dubai where she has no independent residency status and does not wish to reside. The trial judge was aware of the mother’s desire not to return, and to the very real possibility that she might remain in Ontario even if the children were ordered to return. This possibility was a foundational premise of his assessment of the likelihood of serious harm. The trial judge heard the parties and the expert evidence, and he concluded that it would be in the children’s best interests to return to Dubai, even if the mother did not follow them. The trial judge did not rely on the view that the mother’s conduct was the source of serious harm. Nor did he rely on the idea that the mother was obliged to accept the father’s offer to secure her residency status in Dubai. The mother did not demonstrate that his conclusions were unsupported by the evidence or otherwise reflected a palpable and overriding error.

                    The allegations of serious harm faced by the children in the instant case also relate to the mother’s claim that the UAE courts’ parenting decisions are not made according to the best interests of the child. The trial judge was aware that aspects of UAE law are said to conflict with Ontario’s conception of the best interests of children. The division of parental responsibilities on the basis of gender is inconsistent with the gender equality upon which the allocation and exercise of custody and access rights rests in Ontario law. Such a significant discrepancy required the trial judge to determine whether the best interests of the child principle would nevertheless prevail under UAE law should he order the return of the children. The trial judge relied on the testimony of expert witnesses to examine how those rules are applied by UAE courts. He concluded, on the strength of that evidence, that the provisions mandating the allocation of parental responsibilities on the basis of gender are not automatic or imperative, but are rather subject to the discretion of judges who ultimately decide custody and access on the basis of the best interests of the child. The mother has offered no principled basis to revisit the trial judge’s conclusion.

                    In making the return order under s. 40, the judge took into account undertakings made by the father in a settlement offer that would alleviate the precariousness of the mother’s residency status and thereby facilitate her return with the children, should she choose to do so. These undertakings were, for the trial judge, protective measures that promoted the best interests of the children in the event that the mother did decide to return. He considered the undertakings to be adequate. There is no reason to disturb this conclusion, but the undertakings should have been made explicit in the order, given the paramountcy of the best interests of the child principle under s. 40. Accordingly, in dismissing the appeal, it is appropriate to acknowledge that the father is bound by his undertakings.

                    Per Karakatsanis, Brown, Martin and Jamal JJ. (dissenting): There is agreement with the majority’s discussion of the applicable legal principles but disagreement with the application of the law to the instant case. A proper application of the law to the facts establishes that the children would suffer serious harm if removed from Ontario. Therefore, the appeal should be allowed, the order of the trial judge set aside, and the matter returned to a different judge of the Ontario Superior Court to make a parenting order on an expedited basis.

                    While the CLRA aims to discourage child abduction and provides for the return of abducted children, s. 23 recognizes that, in some cases, the objective of discouraging abduction must yield to the paramount objective of preventing serious harm to children. A court’s s. 23 determination is discretionary and generally attracts appellate deference. Appellate courts should not intervene simply because they would have weighed the likelihood or severity of harm differently. But appellate deference is not without limit. An appeal court may intervene if there has been a material error, a serious misapprehension of the evidence, or an error in law.

                    In the instant case, the trial judge made material errors. He seriously misapprehended the evidence in evaluating the likelihood of harm the children would suffer if they are separated from the mother and returned to the father. The likelihood of harm turned on the mother’s claim that she will not return to Dubai. The trial judge was not sure he believed the mother’s claim of non‑return and assigned very little weight to it. He relied on inconsistencies in the mother’s account of tangential and largely irrelevant matters and ignored several crucial relevant considerations supporting the mother’s claim that she would not return, thereby tainting his conclusion on the likelihood of the anticipated harm. Any suggestion that the mother has self‑engineered her claim of serious harm by steadfastly refusing to return to Dubai is rejected. When a parent justifiably refuses to return, the principle against self‑engineered harm does not apply. The mother provided reasonable and legitimate reasons for refusing to return to Dubai. Her precarious residency status in Dubai, her bases for refusing the father’s settlement offer purporting to provide her with benefits if she returns, her legitimate concerns about living under the laws of the UAE as a woman, and her connections in Canada cumulatively rebut any suggestion of self‑engineered harm. The trial judge was required to turn his mind to the relevant factors that go to the believability of the mother’s claim but he ignored relevant evidence as to why she would not return to Dubai. This omission is a material error because the trial judge’s reasons give rise to the reasoned belief that he must have forgotten, ignored, or misconceived the evidence in a way that affected his conclusion. The trial judge’s failure to address crucial relevant considerations shows that he seriously misapprehended the evidence and made an arbitrary decision to place very little weight on the mother’s claim that she would not return to Dubai. The mother established that the children would likely be harmed if returned to Dubai.

                    The trial judge also misapprehended the evidence in evaluating the severity of the harm to the children. The trial judge’s own factual findings regarding the expert evidence and the circumstances of these children demonstrated that the children would suffer serious harm if they lose their mother as their primary caregiver. The trial judge’s contrary conclusion does not attract deference because it contained material errors and failed to address the particular circumstances of the children. Despite the jurisprudence repeatedly recognizing that young children can suffer serious emotional and psychological harm if removed from their primary caregiver, and despite the trial judge accepting expert evidence and taking judicial notice to the same effect, the trial judge found that the children would not suffer serious harm if separated from their mother. This was a serious misapprehension of the evidence and invites appellate intervention. The trial judge’s conclusion gives rise to the reasoned belief that he must have misapprehended the evidence in a way that affected his conclusion. Absent a misapprehension of evidence, the trial judge’s conclusion is inexplicable. The trial judge’s reasons did not apply the case law or the expert evidence to the children’s particular circumstances. In fact, in his s. 23 analysis, the trial judge did not address the children’s circumstances at all. He simply stated bald conclusions.

                    The trial judge did not determine that the children would not suffer serious harm by relying on the mitigating effects of alternate caregivers. The trial judge did not address this factor in his s. 23 analysis. Even if the trial judge had sought to rely on the alternate care proposed by the father to mitigate the harm that the children would suffer if separated from their mother, it is difficult to understand how being cared for by the proposed alternate caregivers could adequately mitigate the harm that the children would suffer if they are separated from their mother.

Cases Cited

By Kasirer J.

                    Applied: Hickey v. Hickey, [1999] 2 S.C.R. 518; Van de Perre v. Edwards, 2001 SCC 60, [2001] 2 S.C.R. 1014; considered: Ojeikere v. Ojeikere, 2018 ONCA 372, 140 O.R. (3d) 561; referred to: Droit de la famille — 3451, [1999] R.D.F. 641; Droit de la famille — 131294, 2013 QCCA 883, [2013] R.J.Q. 849; Thomson v. Thomson, [1994] 3 S.C.R. 551; Office of the Children’s Lawyer v. Balev, 2018 SCC 16, [2018] 1 S.C.R. 398; Droit de la famille — 15751, 2015 QCCA 638; Geliedan v. Rawdah, 2020 ONCA 254, 446 D.L.R. (4th) 440; L.S.I. v. G.P.I., 2011 ONCA 623, 285 O.A.C. 111; E. (H.) v. M. (M.), 2015 ONCA 813, 393 D.L.R. (4th) 267; Young v. Young, [1993] 4 S.C.R. 3; Ontario (Children’s Lawyer) v. Ontario (Information and Privacy Commissioner), 2018 ONCA 559, 141 O.R. (3d) 481; Canadian Foundation for Children, Youth and the Law v. Canada (Attorney General), 2004 SCC 4, [2004] 1 S.C.R. 76; Bolla v. Swart, 2017 ONSC 1488, 92 R.F.L. (7th) 362; W.D.N. v. O.A., 2019 ONCJ 926, 35 R.F.L. (8th) 190; S. (D.M.) v. S. (C.L.), 2016 BCSC 1551, 91 R.F.L. (7th) 202; Onuoha v. Onuoha, 2021 ONSC 2228, 54 R.F.L. (8th) 1, aff’g 2020 ONSC 6849, 49 R.F.L. (8th) 115; Volgemut v. Decristoforo, 2021 ONSC 7382; Ajayi v. Ajayi, 2022 ONSC 5268, 473 D.L.R. (4th) 609; M. (R.A.) v. M. (Y.Y.), 2005 BCPC 259, 48 Imm. L.R. (3d) 301; A. (M.A.) v. E. (D.E.M.), 2020 ONCA 486, 152 O.R. (3d) 81; Leigh v. Rubio, 2022 ONCA 582, 75 R.F.L. (8th) 251; Aldush v. Alani, 2022 ONSC 1536, 74 R.F.L. (8th) 113; Barendregt v. Grebliunas, 2022 SCC 22; B.J.T. v. J.D., 2022 SCC 24; Gordon v. Goertz, [1996] 2 S.C.R. 27; Jamali v. Gillani, 2021 BCSC 2134; Re A. (A Minor) (Abduction), [1988] 1 F.L.R. 365; S.A.G. v. C.D.G., 2009 YKSC 21; Hage v. Bryntwick, 2014 ONSC 4104; P. (J.) v. P. (T.N.), 2016 ABQB 613, 90 R.F.L. (7th) 211; C. v. C. (Minor: Abduction: Rights of Custody Abroad), [1989] 2 All E.R. 465; In re J. (A Child) (Custody Rights: Jurisdiction), [2005] UKHL 40, [2006] 1 A.C. 80; General Motors Acceptance Corp. of Canada v. Town and Country Chrysler Ltd., 2007 ONCA 904, 288 D.L.R. (4th) 74; Hapag‑Lloyd AG v. Iamgold Corp., 2021 FCA 110; Larche v. Ontario (1990), 75 D.L.R. (4th) 377; M.M. v. United States of America, 2015 SCC 62, [2015] 3 S.C.R. 973; Cannock v. Fleguel, 2008 ONCA 758, 303 D.L.R. (4th) 542; R.F. v. M.G., [2002] R.D.F. 785; Brown v. Pulley, 2015 ONCJ 186, 60 R.F.L. (7th) 436.

By Jamal J. (dissenting)

                    Ojeikere v. Ojeikere, 2018 ONCA 372, 140 O.R. (3d) 561; B.J.T. v. J.D., 2022 SCC 24; Van de Perre v. Edwards, 2001 SCC 60, [2001] 2 S.C.R. 1014; Hickey v. Hickey, [1999] 2 S.C.R. 518; Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235; R. v. G.F., 2021 SCC 20; Waxman v. Waxman (2004), 186 O.A.C. 201; Onuoha v. Onuoha, 2021 ONSC 2228, 54 R.F.L. (8th) 1; Gordon v. Goertz, [1996] 2 S.C.R. 27; Young v. Young, [1993] 4 S.C.R. 3; Catholic Children’s Aid Society of Metropolitan Toronto v. M. (C.), [1994] 2 S.C.R. 165; A. (M.A.) v. E. (D.E.M.), 2020 ONCA 486, 152 O.R. (3d) 81; R.J.F. v. C.M.F., 2014 ABCA 165, 575 A.R. 125; Aldush v. Alani, 2022 ONSC 1536, 74 R.F.L. (8th) 113; R. v. Spence, 2005 SCC 71, [2005] 3 S.C.R. 458; R. v. Find, 2001 SCC 32, [2001] 1 S.C.R. 863.

Statutes and Regulations Cited

Canadian Charter of Rights and Freedoms , ss. 2(a) , 7 .

Civil Code of Québec.

Children’s Law Act, 2020, S.S. 2020, c. 2.

Children’s Law Reform Act, R.S.O. 1990, c. C.12 [am. c. 25, Sch. 1], Part III, ss. 18(1), 19, 20(1), 22, 23, 24, 40 to 46, 69.

Extra-Provincial Enforcement of Custody Orders Act, R.S.A. 2000, c. E‑14.

Family Law Act, S.B.C. 2011, c. 25.

Federal Law No. 28 of 2005 on Personal Status (UAE).

Treaties and Other International Instruments

Convention on the Civil Aspects of International Child Abduction, Can. T.S. 1983 No. 35, preamble, Article 13(1)(b).

Authors Cited

Bailey, Martha. “Canada’s Conflicted Approach to International Child Abduction”, in Bill Atkin, ed., The International Survey of Family Law. Bristol: Jordan, 2016, 81.

Bala, Nicholas. “O.C.L. v. Balev: Not an ‘Evisceration’ of the Hague Convention and the International Custody Jurisdiction of the CLRA” (2019), 38 C.F.L.Q. 301.

Bala, Nicholas, and Jacques Chamberland. “Family Violence and Proving ‘Grave Risk’ for Cases Under the Hague Convention Article 13(b)”, Queen’s Law Research Paper No. 91, 2017.

Beaumont, Paul R., and Peter E. McEleavy. The Hague Convention on International Child Abduction. New York: Oxford University Press, 1999.

Chamberland, Jacques. “Country Report — Canada” (2005), 9 Judges’ Newsletter on International Child Protection 75.

Chamberland, Jacques. “Domestic Violence and International Child Abduction: Some Avenues of Reflection” (2005), 10 Judges’ Newsletter on International Child Protection 70.

Eekelaar, John M. “International Child Abduction by Parents” (1982), 32 U.T.L.J. 281.

Gosselain, Caroline. “Child Abduction and Transfrontier Access: Bilateral Conventions and Islamic States”, in Permanent Bureau of the Conference, Preliminary Document No. 7, 2002.

Grammaticaki‑Alexiou, Anastasia. “Best Interests of the Child in Private International Law”, in Collected Courses of the Hague Academy of International Law, vol. 412. Boston: Brill Nijhoff, 2020, 253.

Hague Conference on Private International Law. 1980 Child Abduction Convention — Guide to Good Practice, Part VI, Article 13(1)(b). Hague, 2020.

Hale, Brenda. “Taking Flight — Domestic Violence and Child Abduction” (2017), 70 Current Legal Problems 3.

Schuz, Rhona. The Hague Child Abduction Convention: A Critical Analysis. Oxford: Hart Publishing, 2013.

Schuz, Rhona. “The Relevance of Religious Law and Cultural Considerations in International Child Abduction Disputes” (2010), 12 J.L. & Fam. Stud. 453.

                    APPEAL from a judgment of the Ontario Court of Appeal (Lauwers, Hourigan and Brown JJ.A.), 2021 ONCA 614, 158 O.R. (3d) 481, 464 D.L.R. (4th) 571, 62 R.F.L. (8th) 7, [2021] O.J. No. 4678 (QL), 2021 CarswellOnt 12685 (WL), affirming a decision of Conlan J., 2020 ONSC 7789, 475 C.R.R. (2d) 1, [2020] O.J. No. 5507 (QL), 2020 CarswellOnt 18401 (WL). Appeal dismissed, Karakatsanis, Brown, Martin and Jamal JJ. dissenting.

                    Fareen L. Jamal and Fadwa K. Yehia, for the appellant.

                    Bryan R. G. Smith, Lindsey Love‑Forester and Earl A. Cherniak, K.C., for the respondent.

                    Estée Garfin and Ravi Amarnath, for the intervener the Attorney General of Ontario.

                    Caterina E. Tempesta and Sheena Scott, for the intervener the Office of the Children’s Lawyer.

                    Farrah Hudani and Jessica Luscombe, for the intervener the Defence for Children International‑Canada.

                    Paul‑Erik Veel, for the intervener the Canadian Council of Muslim Women.

The judgment of Wagner C.J. and Moldaver, Côté, Rowe and Kasirer JJ. was delivered by

 

                   Kasirer J. —

I.               Overview

[1]                             The outcome of this appeal turns on whether the Ontario courts should exercise jurisdiction over the merits of a custody dispute involving an international child abduction. The dispute stems from the wrongful retention in Ontario by the appellant, the “Mother”, of two very young children who habitually reside in Dubai, in the United Arab Emirates (“UAE”). The children are retained in the province without the consent of the respondent, the “Father”, who remained in Dubai. What is at issue on appeal is not who, as between the parties, should be awarded the disputed custody rights in respect of the two children but instead which court — the Ontario court or a court in the UAE — should decide the matter.

[2]                             The proceedings commenced with an application for an order under s. 40 of the Children’s Law Reform Act, R.S.O. 1990, c. C.12 (“CLRA”), brought by the left‑behind parent. The Father has asked the Ontario Superior Court of Justice for an order to return the children to the UAE, a country that, unlike Canada, is not a party to the Convention on the Civil Aspects of International Child Abduction, Can. T.S. 1983 No. 35 (“Hague Convention”). The Mother answered that she would not return to Dubai and asked the court to take up jurisdiction over custody on the merits.

[3]                             Where a child who is wrongfully removed to or retained in Ontario habitually resides in a country that is not a party to the Hague Convention, Ontario law provides that, but for exceptional circumstances, courts will refrain from exercising jurisdiction and leave the merits to the foreign jurisdiction with which the child has a closer connection. One such exception is set forth in s. 23 of the CLRA: a court can exercise jurisdiction to make a parenting order where a child is physically present in Ontario and, on a balance of probabilities, the court is convinced that the child would suffer “serious harm” if removed from the province.

[4]                             In opposing the application, the Mother recalls that, like her, the children are Canadian citizens. She says that these two young children would suffer serious harm if separated from her, their primary caregiver. The Mother also argues they would suffer serious harm because, under UAE law, custody would not be decided pursuant to the principle of the best interests of the child as understood in Ontario.

[5]                             After an 11‑day trial, the Ontario court declined jurisdiction because it was not satisfied that the children would suffer serious harm if sent home to Dubai. A majority of the Court of Appeal confirmed the return order. The dissenting judge was of the view that the trial judge had erred in his assessment of serious harm under s. 23 and that the Ontario court should exercise jurisdiction.

[6]                             In this Court, the Mother argues that the trial judge failed to properly identify and apply the best interests of the child principle to the relevant provisions of the CLRA. She says that the return of the young children to Dubai will necessarily separate them from her, their primary caregiver, because her residency status there is wholly dependent on the will and cooperation of the Father. This presents a plain risk of serious harm should these young children be ordered to return. Further, the gender inequality between mothers and fathers that characterizes UAE family law means that custody will not be decided there in accordance with the best interests of the children. By ordering the return of the children, the trial judge failed to apply the best interests principle based on the comprehensive measure set out in s. 24 of the CLRA, which should have led him to refuse the application given the serious harm return would cause to the children.

[7]                             The Mother is right to say that the lower courts were bound to apply the principle of the best interests of the child to the problem at hand and to do so from the child’s‑eye point of view. But she misconstrues, in my respectful view, how the Ontario legislature has directed courts to apply the best interests principle to the question of jurisdiction over children who are wrongfully retained in Ontario.

[8]                             Section 19 sets the objectives for the law relating to decision‑making responsibility and parenting time in Part III of the CLRA, including applications for the return of wrongfully removed children to countries party to the Hague Convention (s. 46(2)) and non‑Hague Convention countries (s. 40). In addition to discouraging child abduction, the legislature seeks to ensure that the child’s best interests are paramount to the making of ultimate parenting orders and that parenting determinations be made in the place to which the child has the closest connection, barring exceptional circumstances.

[9]                             The return order procedure in s. 40 of the CLRA thus starts from the premise that the best interests of the child are aligned with their prompt return to their habitual place of residence so as to minimize the harmful effects of child abduction. Returning the child to the jurisdiction with which they have the closest connection is also understood to be in the child’s best interests. The analysis of the jurisdictional questions contemplated in s. 40, including the risk of serious harm in s. 23, starts from this ordinary alignment of best interests and focuses on factors that would tend to establish, as an exception, serious harm if the child was returned. Contrary to the Mother’s position, the assessment is not a comprehensive comparison of the child’s life in the two jurisdictions or a broad-based best interests test as is conducted for a parenting order on the merits.

[10]                         The approach advocated for by the Mother risks conflating decisions on jurisdiction with custody decisions on the merits and would encourage forum‑shopping in future cases. This would reduce these decisions, as Hourigan J.A. wrote in the Court of Appeal, “to a means for Ontario courts to prefer this province’s system of justice over those of foreign jurisdictions under the guise of child safety” (2021 ONCA 614, 158 O.R. (3d) 481, at para. 79). Worse still, it would invite wrongful abductions to the province for the purpose of grounding jurisdiction there which, as Chamberland J.A. once wrote in a Quebec case, would encourage parents [translation] “to take the law into their own hands and go to another jurisdiction in the hope, whether conscious or not, that the courts there will be more receptive” (Droit de la famille — 3451, [1999] R.D.F. 641 (Que. C.A.), at p. 647, cited with approval in a non‑Hague Convention abduction case in Droit de la famille — 131294, 2013 QCCA 883, [2013] R.J.Q. 849, at para. 46). While the trial judge did not find in this case that the Mother’s disinclination to return was itself the cause of serious harm, it bears recalling that, in both Hague and non‑Hague Convention settings, the courts recognize that a parent should not be allowed to create a situation that is potentially harmful to the child and then rely upon it to establish a risk of harm to the child (see Ojeikere v. Ojeikere, 2018 ONCA 372, 140 O.R. (3d) 561, at para. 91). Finally, the Mother’s characterization of UAE law as an inherent source of serious harm must be rejected. Expert evidence accepted by the trial judge led him to conclude that the best interests principle would apply to the ultimate custody decision before a UAE court. As Hourigan J.A. observed, “the rather provincial view that unless Ontario law is applied, children will suffer serious harm” could have the unwitting effect of turning Ontario into a haven for child abduction (para. 83; see also para. 136, per Brown J.A., concurring).

[11]                         The trial judge was called upon to decide the fact‑specific, highly individualized question as to whether these two children would suffer serious harm if removed from Ontario. He understood that the separation of children from their primary caregiver typically gives rise to emotional distress for very young children. But he found, on the basis of the evidence, that this distress did not rise to the higher level of serious harm. Based on expert evidence relating to the differences between Ontario and UAE law, the judge determined the best interests of the two children will be the paramount consideration for determining custody in a UAE court. In the absence of a reviewable error, his decision that the serious harm threshold was not met is entitled to deference, as explained by the majority of the Court of Appeal. The custody dispute — undecided here — should be resolved by the courts in the UAE, where the children have their closest connection.

[12]                         Finally, I note the allegations of serious harm faced by the children in this case do not relate to domestic violence or abuse of any kind, but rather to the risk that the Mother, as primary caregiver, will be separated from the children should they be ordered home to Dubai where she has no independent residency status and does not wish to reside. The trial judge was aware of the risk of separation due to the precariousness of the Mother’s residency status. In making the return order, the judge took into account undertakings made by the Father in a “with prejudice” settlement offer that would alleviate that precariousness and thereby facilitate the Mother’s return with the children.

[13]                         Insofar as the Mother’s residency status presents a risk that she will be separated from the children, the undertakings would remove an obstacle to her return to Dubai, should she wish to do so. While undertakings made by a left-behind parent before the Ontario courts may present problems of enforceability before foreign courts, they are well-known and relied upon protective measures in international abduction cases around the world, including in this Court (Thomson v. Thomson, [1994] 3 S.C.R. 551, at p. 599). Here, the undertakings that are directed at easing the Mother’s precarious residency status were, for the trial judge, protective measures that promoted the best interests of the children in the event that the Mother did decide to return. But at the end of the day, the trial judge found that whether or not the Mother returned, and independently of the undertakings, sending the children back home to Dubai would not cause them serious harm within the meaning of s. 23 and would be in their best interests. In this case, the judge did not rely on the view that the Mother’s conduct was the source of serious harm. Nor did he rely on the idea that the Mother was obliged to accept the Father’s offer to secure her residency status in Dubai.

[14]                         The terms of the Father’s offer also do more than just remove potential barriers to the Mother’s return. As Hourigan J.A. observed, the Father has also promised that the Mother will remain the primary residential parent, and that major parenting decisions will be made jointly by the two parents (para. 16). Further, the proposed settlement agreement would — if the Mother accepted it — see her retain custody of the children until they are 18, whether or not she chooses to remarry (A.R., vol. XI, at pp. 310‑11, cls. 3/1 and 3/2). The trial judge found that these promises could be made part of an enforceable order in the UAE courts.

[15]                         When the trial judge asked counsel to make submissions on how the Father’s undertakings could be incorporated into his order, the Mother did not follow up on the invitation. The Father has renewed his settlement offer before this Court. In her written argument, the Mother acknowledges the viability of protective measures to address her circumstances when she states that the trial judge should have required more forceful undertakings to answer the uncertainties surrounding her status in Dubai. Given that the judge took the Father’s undertakings into account but did not expressly record them in the order, I think it prudent to note that the order recognizes that the undertakings would assist the Mother should she wish to return to Dubai with the children. I would thus propose to confirm the return order and, in dismissing the appeal, acknowledge the undertakings relied upon in first instance.

II.            Background

[16]                         The Father is a citizen of Pakistan. He has been living in Dubai since 2008. The Mother moved with her family from Pakistan to Ontario in 2005, where she lived until 2012. She is a citizen of Pakistan and of Canada. In 2012, the Mother and the Father were married in Pakistan and then moved to Dubai, where the Father works in the banking sector. While the Mother has worked sporadically in Dubai, she does not have an independent residency status there. She has been sponsored by the Father throughout their marriage.

[17]                         The parties have two children, a girl, Z. (born in 2016, aged four at trial) and a boy, E. (born in 2019, who was about one at trial). Both children are Canadian citizens but, until they were brought to Ontario in June 2020, they had always been resident in Dubai. The Mother is their primary caregiver. Mary, the family’s live-in nanny, contributed significantly to their upbringing. In Dubai, the family lived comfortably. Their daughter Z. attended a well-regarded international school (2020 ONSC 7789, 475 C.R.R. (2d) 1, at para. 30).

[18]                         The parties experienced persistent marital difficulties.

[19]                         On June 19, 2020, the Mother travelled to Ontario with Z. and E. They left Dubai with return air tickets, ostensibly for a trip to visit her family. The Father, who remained in Dubai, agreed to that trip. On or about July 2, 2020, the Mother informed the Father that she intended to stay in Ontario with the children and not return to Dubai. The Father did not consent to the children remaining in Ontario and the Mother did not seek the authorization of a court to retain the children there.

[20]                         Shortly after learning of the Mother’s intentions, the Father initiated proceedings in Ontario seeking an order under s. 40 of the CLRA for the return of the two children to Dubai. He argued that the court should not take up jurisdiction to make a parenting order and that it was in the best interests of the children that all issues of custody and access be decided in the UAE.

[21]                         In response, the Mother said that custody and access should be decided in Ontario under the CLRA. She claimed that it was in the best interests of the children to remain in Ontario with her. The Mother filed submissions requesting sole custody. She depicted the Father as ill-tempered and aggressive and asked that he be granted limited access to the children (trial reasons, at paras. 20 and 70). The Mother also challenged the constitutionality of s. 40 para. 3 of the CLRA, arguing that it was ultra vires the legislative authority of the province and offended various guarantees under the Canadian Charter of Rights and Freedoms .

[22]                         In July 2020, the Father initiated divorce proceedings in Dubai, in which the Mother did not participate. In March 2021, the Father obtained a divorce in Dubai. Under UAE law, divorced non‑nationals are granted a one-year grace period during which they can remain in the country without a residency permit. The Mother’s grace period had expired by the time the Court heard this appeal in April 2022.

[23]                         Prior to the hearing of the s. 40 application, the Father presented a “with prejudice” settlement offer to the Mother’s counsel and filed it in court (trial reasons, at paras. 48‑49). He made a series of undertakings that could be included in the s. 40 order “as a condition of [the Mother] returning” to Dubai with the children (para. 49). In the offer, he undertook to ensure her independent residency in Dubai, in particular, by purchasing property for her in her name. He further undertook that the children reside primarily with her and that major decisions regarding them be made jointly. He wrote that the Mother would be free, should she wish, to contest the proposed custody and financial arrangements when the matter proceeded before the UAE court (see para. 49). The Father also incorporated by reference terms from a draft settlement agreement that he would propose to the UAE court (para. 49). In that draft settlement, the Father proposed to ensure that the Mother would retain custody of the children until they turned 18, whether or not the Mother chose to remarry. The Father stated that these terms could be included in a consent judgment issued by a UAE court. The Mother did not respond to this offer.

[24]                         At trial, both parties testified. The Father called an expert on family law in the UAE, Diana Hamade. The Mother called two experts: Elena Schildgen, an expert on family law in the UAE, and Carol‑Jane Parker, a registered psychotherapist, who provided evidence on the potential impact of separating young children from their primary caregivers.

[25]                         The trial judge had to resolve two issues: first, whether he should exercise jurisdiction for a parenting order on the merits; and second, if the court did not take up jurisdiction, whether to order the return of the children to Dubai pursuant to s. 40 of the CLRA.

III.         Proceedings Below

A.           Superior Court of Justice, 2020 ONSC 7789, 475 C.R.R. (2d) 1 (Conlan J.)

[26]                         The trial judge found that both the Mother and the Father were good, loving and caring parents, who never mistreated, abused or neglected their children (paras. 15 and 480). The Mother has always been the primary caregiver; Mary, the nanny, had helped substantially with the care of Z. and E. (paras. 262 and 291).

[27]                         As witnesses, the Father was “significantly more credible” than the Mother, who “was not a credible witness” (paras. 255‑56). The Mother’s evidence at trial was incompatible with the final order she was asking the court to make. On the one hand, she testified that she supported co-parenting — even joint custody — and, at a minimum, “lots of access” for the Father to the children (para. 281). However, her draft final order sought what the judge described as “anything but lots of access” for the Father (paras. 281‑82). The trial judge also noted inconsistencies in the Mother’s testimony (paras. 257‑74 and 277‑82). He did not accept the Mother’s claims of incidents of physical aggression (paras. 272‑74), or that she suffered religious discrimination and social isolation (para. 369), or her downplaying of the role that the long serving nanny, Mary, had with the children in Dubai.

[28]                         The trial judge accepted the entirety of the expert evidence of the Father, Ms. Hamade, who was “very experienced in the courts in the United Arab Emirates” and “fluent in the Arabic language in which the original relevant legislation was written” (para. 297). He accepted only a limited amount of Ms. Parker’s evidence, and some of the opinions expressed by the Mother’s expert, Ms. Schildgen, whose testimony on the state of UAE law was “devastating” for the Mother’s case (para. 293). The cross‑examination of Ms. Schildgen “eviscerated the very underpinning” of the Mother’s case, as it established that the best interests of the children would be the paramount consideration in any decision made regarding them in Dubai (para. 304).

[29]                         The trial judge assessed the entirety of the testimony he had heard. On the basis of that assessment, he made “four key findings” from the expert evidence (para. 294). On the foreign law, the best interests of Z. and E. would be the paramount consideration in determining custody in Dubai, and the settlement proposed by the Father, if agreed to by the Mother, could be incorporated into a valid court order in Dubai and would be enforceable. As to the consequences of a return of the children, he said the expert opinion established that children may face emotional and psychological repercussions from being separated from their primary caregiver. But as to what would occur with Z. and E., if separated from the Mother, the trial judge said that, based on the admissible evidence given by Ms. Parker, that specific effect remained unknown.

[30]                         After having reviewed the applicable legal principles, the trial judge determined that he should not assume jurisdiction under ss. 22 or 23 of the CLRA or under his parens patriae jurisdiction. He summarized his own findings in respect of the “serious harm” exception in s. 23 as follows:

(i)     there is no evidence at trial that Z. and E. are in any risk of being physically harmed if they return to Dubai;

(ii)    there is some circumstantial evidence at trial (through Ms. Parker and her opinions about infants, generally) that Z. and E. could be at risk of emotional and psychological harm if they are returned to Dubai without [the Mother];

(iii)   there is no evidence at trial about the views and preferences of the children;

(iv)   there is a claim by [the Mother] that she will not return to Dubai if the children are ordered to return there;

(v)    there is nothing else in the evidence at trial that this [c]ourt finds to be relevant to the serious harm assessment for Z. and E. . . .; and

(vi)   more specifically, considering the evidence of Ms. Hamade and Ms. Schildgen, there is a total absence of any reliable evidence at trial that the court system in Dubai will do anything other than (a) determine custody in accordance with the best interests of Z. and E., if contested; and (b) award custody to [the Mother], if contested, and (c) approve the settlement proposal tendered by [the Father], if agreed to by the [M]other. [Emphasis in original; para. 366.]

[31]                         He rejected the Mother’s argument that the children would suffer serious harm if they were returned because a court in Dubai would not “really employ a best interests of the child analysis” (para. 367 (emphasis deleted)). He placed “very little weight” on the assertion by the Mother that she would not return to Dubai, citing his earlier conclusions on credibility (para. 368).

[32]                         The trial judge accorded significant importance to the Father’s “with prejudice” offer (paras. 48‑49) and was of the view that Dubai courts would approve and enforce the settlement proposal tendered by the Father, if agreed to by the Mother (para. 366(vi); see also paras. 294 and 301). His conclusion was based on the expert evidence of Ms. Hamade, who, contrary to the Mother’s expert, was “intimately familiar with all aspects of [the Father’s] proposed settlement agreement” (para. 297; see also paras. 193‑94).

[33]                         While the trial judge recognized that young children, like Z. and E., can face negative emotional effects when separated from their primary caregiver (paras. 305 and 366(ii)), he determined that the anticipated risk did not rise to the required level of “serious harm”. The trial judge concluded that he could not exercise jurisdiction under s. 23 of the CLRA because he was not satisfied that Z. and E. would suffer serious harm if they were removed from Ontario (para. 370).

[34]                         Turning to the issue of whether the children should be returned to Dubai pursuant to s. 40 of the CLRA, he concluded that it was in the best interests of the children to do so, with or without the Mother (paras. 381 and 387). Recognizing that s. 40 provided him with broad discretionary powers and that he was not bound to order the return of the children, the trial judge wrote that it was the “only appropriate order” (para. 384). Dubai was their real home.

[35]                         The trial judge also dismissed the Mother’s challenge to the constitutional validity of s. 40 para. 3 of the CLRA (paras. 463‑64).

[36]                         He declared that the Mother had wrongfully retained the children in Ontario and ordered that they be returned to Dubai (para. 469). Lastly, before the order was finalized, and before the children would return to Dubai, he offered all parties the opportunity to make further submissions on whether to include the Father’s settlement proposal in the order (para. 472). The Mother did not avail herself of this opportunity.

B.            Court of Appeal, 2021 ONCA 614, 158 O.R. (3d) 481 (Lauwers, Hourigan and Brown JJ.A.)

[37]                         A majority of the court dismissed the appeal, confirming the trial judge’s order to return the children to Dubai. Hourigan J.A. wrote reasons on the jurisdictional issues. Brown J.A. wrote principally on the constitutional matters, which are not at issue before this Court. In dissent, Lauwers J.A. concluded the Superior Court of Justice had jurisdiction to make a parenting order based on s. 23 of the CLRA.

(1)        Hourigan J.A. (Brown J.A. Concurring)

[38]                         Recalling the applicable standard of review, Hourigan J.A. observed that the trial judge’s decision on serious harm was subject to deference on appeal. He dismissed the Mother’s argument on s. 23 as an inappropriate invitation to reweigh evidence settled at trial. The trial judge was aware of the Mother’s precarious residency status and the resulting risk of separation. For Hourigan J.A., the trial judge was entitled to rely on the expert evidence and conclude that there were workable solutions for the Mother to obtain her residency, noting that the Mother declined to make submissions on the Father’s settlement offer. The trial judge considered the expert evidence and it was open to him to conclude that an agreement between the parties could be incorporated in a court order (C.A. reasons, at paras. 62‑63). He disagreed with the dissenting judge’s view that the trial judge failed to consider the possibility that the Father would renege on the offer and seek to limit the Mother’s access to the children. “Perhaps the one indisputable fact established on the record”, wrote Hourigan J.A., “was that the [Mother], not the [Father], sought to limit access to the children” (para. 71).

[39]                         Moreover, the trial judge was aware of the inconsistencies between UAE and Canadian law, but properly focussed on whether these discrepancies would cause serious harm to the children. No reviewable error was shown. To say that the discrepancies did ground jurisdiction in Ontario would send a message that parents in the UAE could remove children to the province and not be required to send them home (para. 83).

[40]                         Hourigan J.A. then turned to the trial judge’s assessment of whether the separation of the children from their primary caregiver presented a risk of serious harm. Based on the expert evidence, the trial judge could fairly conclude that the children might be at risk of emotional or psychological damage if separated from the Mother, but that the threshold for serious harm had not been met (para. 91). It would be mistaken to assume that whenever young children are separated from their primary caregiver there will always be a risk of serious harm, regardless of the circumstances of each case (paras. 92‑94). This would serve to encourage child abduction, unduly focus the analysis on the preferences of the custodial parent (rather than the best interests of the child) and undermine the individualized nature of the serious harm inquiry. Again, Hourigan J.A. found no basis for disturbing the trial judge’s conclusion.

(2)        Brown J.A. (Hourigan J.A. Concurring)

[41]                         In concurring reasons, Brown J.A. dismissed the Mother’s grounds of appeal regarding the constitutional challenge. Although this Court did not grant the Mother’s motion for leave to argue the constitutional issues, Brown J.A.’s explanation of the purposes of s. 40 of the CLRA is useful to the jurisdictional questions that remain in dispute.

[42]                         Brown J.A. observed that return orders made under s. 40 para. 3 share the same purposes as those made under the Hague Convention, that is, “to protect a child from the harmful effects of their wrongful removal or retention and to return a child wrongfully removed or retained to the jurisdiction which is most appropriate for the determination of custody and access” (para. 131). Brown J.A. also noted that the scope of a best interests inquiry under s. 40 para. 3 differs from that undertaken when custody is at issue under s. 24. When determining whether to return children to a different jurisdiction, a court must also consider the CLRA’s other policy objectives, such as discouraging the abduction of children and avoiding concurrent exercises of jurisdiction (para. 186). On the facts of this case, he was satisfied that the trial judge performed the multi-factored analysis required by s. 40 para. 3 of the CLRA, and that his findings of fact supported the conclusion that the return of the children to Dubai was in their best interests (paras. 189‑90).

(3)        Lauwers J.A. (Dissenting)

[43]                         For Lauwers J.A., the trial judge erred, both in law and in fact, when he concluded that the return of the children would not expose them to serious harm. He would have applied the exception in s. 23, set aside the return order and directed the Ontario Superior Court take up jurisdiction to make a parenting order.

[44]                         In this case, the children will be separated from their Mother because of her precarious residency status in Dubai. The separation of young children from their primary caregiver constitutes, in itself, a risk of serious harm (para. 291). The trial judge committed a palpable and overriding error when he wrote that the impact of the involuntary separation was “unknown”. This was precisely the issue the judge had to determine under s. 23 (para. 287).

[45]                         The Father’s contingent undertakings do not effectively mitigate the Mother’s precariousness in Dubai where she has no independent legal right to reside (paras. 297‑302). The trial judge failed to consider whether the undertakings would be enforceable in Dubai, and he could not rely on the Father’s credibility at trial to presume that he will honour the proposals he made (para. 302). If the Father was to rescind the proposal, UAE law would govern the dispute between the parties. The automatic assignment of decision‑making on the basis of gender under UAE law is a pronounced departure from Ontario’s understanding of the best interests of the child. Together, these interrelated factors amount to serious harm for the children (paras. 312‑18).

IV.         Issues

[46]                         The Mother’s appeal raises two principal issues.

[47]                         First, she says the trial judge erred by declining to take up jurisdiction of the custody dispute on the merits. He wrongly concluded that jurisdiction could not be founded on s. 23 of the CLRA. Specifically, he erred in ignoring factors that, if properly considered, would have satisfied the test for serious harm.

[48]                         Second, she says the trial judge erred in ordering that the children should be returned to Dubai pursuant to s. 40 para. 3 of the CLRA. He wrongly rejected a broad-based best interests analysis that should guide the decision whether or not to issue a return order. Further, the trial judge should have made an interim parenting order maintaining the children in her care until a mechanism was put in place to assure, in particular, that property was purchased in her name to guarantee her residency status.

V.           Analysis

A.           Child Abduction in Ontario and Jurisdictional Issues

[49]                          Whether or not the dispute involves a country that is a party to the Hague Convention, when a child is wrongfully removed from their habitual residence and brought to Ontario, courts will, as a general rule, decline to exercise jurisdiction over the merits of the custody dispute and order that the child be returned home. This reflects a legislative policy expressed in s. 19 for the whole of Part III of the CLRA. Through this policy, the legislature seeks to discourage child abductions and the wrongful removal and retention of children to Ontario as well as to ensure that parenting matters are determined by the jurisdiction to which the child has a closer connection.

[50]                          Ontario sees the policy of discouraging child abduction as aligned with the principle of the best interests of the child. This finds expression in the province’s acceptance of Canada’s adherence to the Hague Convention, pursuant to s. 46(2) of the CLRA. The Hague Convention endeavours to protect children from the harmful effects of removal or retention of a child in breach of the left‑behind parent’s custody rights and also seeks to establish procedures to ensure the prompt return of children to the country of their habitual residence (see, e.g., Thomson, at pp. 575‑76; Office of the Children’s Lawyer v. Balev, 2018 SCC 16, [2018] 1 S.C.R. 398, at paras. 22‑24; Droit de la famille — 15751, 2015 QCCA 638, at para. 17 (CanLII)). Parties to the Hague Convention recognize that they share the legal principle that “the interests of children are of paramount importance in matters relating to their custody” (Hague Convention, preamble; see also Geliedan v. Rawdah, 2020 ONCA 254, 446 D.L.R. (4th) 440, at para. 37).

[51]                          When children are wrongfully removed from non‑party states and brought to Ontario or another Canadian province or territory, the Hague Convention does not apply. To address international abductions falling outside the scope of the Hague Convention, Canadian provinces and territories have enacted various statutory regimes (see, e.g., Family Law Act, S.B.C. 2011, c. 25 (British Columbia); Extra-Provincial Enforcement of Custody Orders Act, R.S.A. 2000, c. E‑14 (Alberta); The Children’s Law Act, 2020, S.S. 2020, c. 2 (Saskatchewan); on the application of the Civil Code of Québec to non‑Hague Convention cases, see Droit de la famille — 131294). In Ontario, the relevant rules are found in Part III of the CLRA.

[52]                          A review of legislation in this area reveals that, in general, Canadian provinces treat child abductions in non‑Hague Convention cases in a manner methodologically comparable to the Convention: first, by declining to decide parental disputes on the merits with respect to children who do not habitually reside in the province or territory, and second, by favouring the return of children to the jurisdiction of their habitual residence. However, these similarities do not mean that an application brought under provincial legislation is treated the same way as one brought subject to the rules of the Hague Convention (Geliedan, at paras. 26‑34).

[53]                          In Thomson, a Hague Convention case, this Court clarified that the legislation governing non‑Hague and Hague Convention disputes “operate independently of one another” (p. 603; see also L.S.I. v. G.P.I., 2011 ONCA 623, 285 O.A.C. 111, at para. 46). As pointed out by Laskin J.A. in Ojeikere, in Hague Convention cases, “Ontario courts can have confidence that whatever jurisdiction decides on a child’s custody it will do so on the basis of the child’s best interests”, but they cannot have the same confidence in cases involving non‑party jurisdictions (para. 60; see also Geliedan, at paras. 37‑38 and 45). The Ontario legislature makes plain, in s. 19(a) of the CLRA, that the ultimate determination of a parenting order on the merits will be made on the basis of the best interests of the child. To account for the fact that, in the non‑Hague Convention context, Ontario courts do not benefit from the a priori assumption that the best interests of the child principle will be applied to the merits of the custody dispute in the foreign country, judges assessing petitions for return to non‑party jurisdictions must therefore consider the tenor of foreign law, generally through expert evidence adduced by the parties. Nevertheless, in Thomson, La Forest J. explained that it is not improper to look at the Hague Convention for the interpretation of domestic legislation, “since the legislature’s adoption of the Convention is indicative of the legislature’s judgment that international child custody disputes are best resolved by returning the child to its habitual place of residence” (p. 603; for the CLRA, see N. Bala, “O.C.L. v. Balev: Not an ‘Evisceration’ of the Hague Convention and the International Custody Jurisdiction of the CLRA” (2019), 38 C.F.L.Q. 301, at p. 308).

[54]                          The children in this case have been removed from the UAE, a state which is not a party to the Hague Convention. It follows that this matter should be resolved on the basis of the general provisions of Part III of the CLRA.

B.            The Statutory Scheme in Ontario

[55]                         Part III of the CLRA sets out the provisions concerning orders for decision‑making responsibility and parenting time made by Ontario courts under the law of Ontario.^[1] An Ontario court will not make a “parenting order” unless it has jurisdiction to do so. Section 22 concerns jurisdiction over a child habitually resident in Ontario or a child physically present in Ontario where, in particular, substantial evidence concerning the best interests of the child is available in the province. Moreover, a court may exercise its parens patriae jurisdiction, preserved by s. 69.

[56]                          An Ontario court can also exercise its jurisdiction if the child would suffer serious harm as a result of being removed from Ontario. On appeal to this Court, the Mother relies only on s. 23 as the basis for the Ontario court’s jurisdiction:

23 Despite sections 22 and 41, a court may exercise its jurisdiction to make or to vary a parenting order or contact order with respect to a child if,

 

(a) the child is physically present in Ontario; and

 

(b) the court is satisfied that the child would, on the balance of probabilities, suffer serious harm if,

 

(i) the child remains with a person legally entitled to decision‑making responsibility with respect to the child,

 

(ii) the child is returned to a person legally entitled to decision‑making responsibility with respect to the child, or

 

(iii) the child is removed from Ontario.

[57]                          Within Part III, ss. 40 to 46 of the CLRA bear on “Decision-Making Responsibility, Parenting Time and Contact — Extra-Provincial Matters”. Section 40, the provision upon which the Father relies in this case, applies to the application for the return of children wrongfully removed to or retained in Ontario when those children are not subject to the Hague Convention. It provides:

40 Upon application, a court,

 

(a) that is satisfied that a child has been wrongfully removed to or is being wrongfully retained in Ontario; or

 

(b) that may not exercise jurisdiction under section 22 or that has declined jurisdiction under section 25 or 42,

 

may do any one or more of the following:

 

1. Make such interim parenting order or contact order as the court considers is in the best interests of the child.

 

2. Stay the application subject to,

 

i. the condition that a party to the application promptly commence a similar proceeding before an extra-provincial tribunal, or

 

ii. such other conditions as the court considers appropriate.

 

3. Order a party to return the child to such place as the court considers appropriate and, in the discretion of the court, order payment of the cost of the reasonable travel and other expenses of the child and any parties to or witnesses at the hearing of the application.

[58]                          Finally, s. 19 states the purposes of the whole of Part III — including the rules in s. 40 and the rules on jurisdiction, such as s. 23:

19 The purposes of this Part are,

 

(a) to ensure that applications to the courts respecting decision‑making responsibility, parenting time, contact and guardianship with respect to children will be determined on the basis of the best interests of the children;

 

(b) to recognize that the concurrent exercise of jurisdiction by judicial tribunals of more than one province, territory or state in relation to the determination of decision‑making responsibility with respect to the same child ought to be avoided, and to make provision so that the courts of Ontario will, unless there are exceptional circumstances, refrain from exercising or decline jurisdiction in cases where it is more appropriate for the matter to be determined by a tribunal having jurisdiction in another place with which the child has a closer connection;

 

(c) to discourage the abduction of children as an alternative to the determination of decision‑making responsibility by due process; and

 

(d) to provide for the more effective enforcement of parenting orders and contact orders, and for the recognition and enforcement of orders made outside Ontario that grant decision‑making responsibility, parenting time or contact with respect to a child.

[59]                          In sum, parents whose children have been abducted from a non‑party country can apply for their return pursuant to s. 40 of the CLRA. Unless the abducting parent demonstrates that Ontario courts should make parenting orders on any one of the four bases outlined above (ss. 22(1)(a) or (b) or 23, or parens patriae jurisdiction), the courts should decline to exercise jurisdiction with respect to a child (Ojeikere, at para. 12; E. (H.) v. M. (M.), 2015 ONCA 813, 393 D.L.R. (4th) 267, at paras. 22‑26). In this case, all that remains in dispute before this Court is whether s. 23 applies such that the Ontario courts should take jurisdiction over the merits of the dispute and, accordingly, refuse the Father’s application for the return of the children to Dubai.

C.            Jurisdiction and the Best Interests of the Child

[60]                          This case invites the Court to clarify the role the best interests of the child principle plays in the interpretation and application of the CLRA’s jurisdictional rules. The Mother argues that where jurisdiction is premised on the “serious harm” exception, judges must undertake a broad-based best interests analysis, having regard to the factors set out in s. 24(3) (A.F., at para. 10). The Father answers that s. 23 already reflects an overriding concern for the best interests of children. He says that courts should not embark on a broad-based best interests analysis like the one conducted when deciding custody on the merits (R.F., at paras. 97‑99).

[61]                          The Mother is not mistaken to point to the importance of the best interests of the child principle. As a general rule in Canadian family law, it is undoubtedly the case that the best interests of the children are the paramount consideration for all decisions that concern children and that best interests are measured from the child’s perspective (see, e.g., Young v. Young, [1993] 4 S.C.R. 3; Ontario (Children’s Lawyer) v. Ontario (Information and Privacy Commissioner), 2018 ONCA 559, 141 O.R. (3d) 481, at para. 58; Canadian Foundation for Children, Youth and the Law v. Canada (Attorney General), 2004 SCC 4, [2004] 1 S.C.R. 76, at para. 9). This is no less true in matters of international abduction, whatever the child’s country of origin, and whether or not the Hague Convention governs the dispute.

[62]                          That said, the Mother’s submissions misapprehend the manner in which the CLRA would have the best interests principle apply to ss. 40 and 23. The legislature’s overriding concern for upholding the best interests of the child principle is encoded in the rules of the CLRA (Ojeikere, at para. 17). Under the statutory scheme, the best interests test calls for a “differing application” in questions of jurisdiction and determinations on the merits (C.A. reasons, at para. 187, per Brown J.A.). On the one hand, when Ontario courts are called upon to make parenting or contact orders on the merits, the best interests of the child are to be comprehensively assessed in light of the factors listed in s. 24 of the CLRA. On the other hand, when courts determine whether Ontario should decline jurisdiction in favour of foreign courts and return children, they are deciding, fundamentally, which court will decide on custody, not custody itself. I agree with Brown J.A.’s explanation: “[the children’s] immediate legal interests concern where the issue of their custody should be determined, not who should have custody”, he wrote, and “[t]he best interests test must be applied in that context” (para. 187).

[63]                          As counsel for the Attorney General of Ontario explain in their helpful factum, the CLRA presumes, following an abduction, that the child’s best interests are aligned with their prompt return to the jurisdiction of their habitual residence unless there are exceptional circumstances that justify Ontario courts taking up jurisdiction (I.F., at para. 6).

[64]                          The premise that the children’s best interests are favoured by their timely return to their home jurisdiction is sound. Child abductions harm children (Balev, at paras. 23‑25; Ojeikere, at para. 16; A. Grammaticaki‑Alexiou, “Best Interests of the Child in Private International Law”, in Collected Courses of the Hague Academy of International Law (2020), vol. 412, 253, at p. 325). As McLachlin C.J. explained in Balev, “[t]he children are removed from their home environments and often from contact with the other parents. They may be transplanted into a culture with which they have no prior ties, with different social structures, school systems, and sometimes languages. Dueling custody battles waged in different countries may follow, delaying resolution of custody issues. None of this is good for children or parents” (para. 23). Moreover, resolving parenting issues in the children’s home jurisdiction fosters stability, while ensuring that custody will be determined by the authorities of the place with which the child has the closer connection, which is an objective set out under s. 19 of the CLRA. Indeed, the jurisdiction from which the children have been removed is usually in the best position to determine which arrangement will be in their best interests (Bolla v. Swart, 2017 ONSC 1488, 92 R.F.L. (7th) 362, at para. 38; W.D.N. v. O.A., 2019 ONCJ 926, 35 R.F.L. (8th) 190, at para. 51; Droit de la famille — 131294, at para. 110). This is explained by the fact that “the courts of the child’s State of habitual residence . . . generally will have fuller and easier access to the information and evidence” relevant to making a “comprehensive best interests’ assessment” (Hague Conference on Private International Law, 1980 Child Abduction Convention — Guide to Good Practice, Part VI, Article 13(1)(b) (2020) (“Guide”), at para. 15; see also J. M. Eekelaar, “International Child Abduction by Parents” (1982), 32 U.T.L.J. 281, at p. 301).

[65]                          Consequently, at the preliminary stage of deciding jurisdiction, it is not the role of the judge to conduct a broad-based best interests inquiry, as they would on the merits of a custody application. Instead, the judge must determine if the court should exercise jurisdiction on the basis of any of the four grounds listed above. If one of the four grounds is made out, the presumption that it is in the children’s best interests to decline jurisdiction and order their return will be rebutted. One such exceptional circumstance is the s. 23 basis for jurisdiction, where there is a risk of serious harm should the child be removed from Ontario.

[66]                          Under s. 23 of the CLRA, the court may exercise its jurisdiction if the abducting parent proves, on a balance of probabilities, that the child would suffer serious harm if removed from Ontario. If serious harm is established, it will be in the child’s best interests for the Ontario court to assume jurisdiction and decide on custody on the merits. Simply put, the presumption in favour of the jurisdiction of habitual residence must give way to the imperative of protecting a child when serious harm is made out.

[67]                          Thus, when deciding whether to exercise jurisdiction under s. 23, judges should not conduct a broad-based best interests analysis, but should rather conduct an individualized assessment of the risk of serious harm. This is also in keeping with the expeditious character of the proceedings associated with an application for a return order which, as this Court noted in Balev, serves to reduce the harmful effects of removal upon the child and the left‑behind parent (paras. 25‑27 and 88‑89). I agree with the Attorney General of Ontario that replacing the serious harm threshold with a broad‑based best interests assessment would conflate the question of jurisdiction with the making of parenting orders and inappropriately dilute the threshold of serious harm (I.F., at paras. 3 and 17). In so doing, it would risk removing “the presumption of a timely return, resulting in Ontario courts exercising jurisdiction in most cases of wrongful abduction or retention” (I.F., at para. 3). In this sense, the Mother’s argument that the s. 23 analysis requires a broad‑based best interests analysis ignores the fundamental distinction between jurisdictional issues and determinations on the merits. It ultimately undermines the purpose of the serious harm exception, that is, to ensure decisions on the merits are made by the appropriate authority in accordance with the best interests of the children (see s. 19(a) and (b)).

[68]                          There is, of course, no doubt that an individualized serious harm analysis may overlap with a full best interests analysis. For instance, many of the best interests factors enumerated under s. 24(3) of the CLRA, which are to be applied when “making a parenting order or contact order”, may also inform a serious harm inquiry, depending on the circumstances of a given case (see, e.g., Ojeikere, at para. 107, per Miller J.A., concurring). However, the best interests factors provided under s. 24 of the CLRA will only be relevant to the extent that they contribute to establishing serious harm, contrary to the broad‑based exercise that s. 24 contemplates for “making a parenting order or contact order” (s. 24(1)). For example, the presence of “family violence”, a factor listed under s. 24(3)(j), could well be relevant to establish serious harm under s. 23 (see, e.g., S. (D.M.) v. S. (C.L.), 2016 BCSC 1551, 91 R.F.L. (7th) 202, at paras. 51‑53; E. (H.), at paras. 122‑25). By contrast, a court may feel that it would be in the child’s best interests that they attend a highly reputable school, but this factor would be unlikely to have weight in a serious harm inquiry.

D.           Serious Harm

(1)          Defining the Scope of the Serious Harm Exception

[69]                         The onus to prove that the child would suffer serious harm rests on the abducting parent (Onuoha v. Onuoha, 2021 ONSC 2228, 54 R.F.L. (8th) 1, at para. 23, aff’g 2020 ONSC 6849, 49 R.F.L. (8th) 115). The burden is demanding. It is not enough to identify a “serious risk” of harm: the court must be satisfied, on a balance of probabilities, that the harm itself would be serious in nature. This high threshold is necessary to ensure that the legislative purposes of discouraging child abductions and avoiding concurrent jurisdiction are properly served. It is not enough to conclude that the return would have a negative impact on the child.

[70]                         Serious harm inquiries are child‑centered. As La Forest J. wrote in Thomson, building upon this Court’s decision in Young, “from a child centered perspective, harm is harm. If the harm were severe enough to meet the stringent test of the Convention, it would be irrelevant from whence it came” (p. 597). One implication of this focus is that, as a general rule, return orders should not become a vehicle for punishing the abducting parent. The relevance of the abducting parent’s conduct is assessed from the child’s point of view.

[71]                         When conducting their s. 23 analysis, judges should consider both the likelihood and the severity of the anticipated harm (Ojeikere, at para. 62). The measure of s. 23 is highly factual and, as the Court of Appeal pointed out, is discretionary, in the sense that it involves the weighing of various factors (para. 52; Ojeikere, at para. 63; E. (H.), at para. 29; Volgemut v. Decristoforo, 2021 ONSC 7382, at para. 99 (CanLII); Ajayi v. Ajayi, 2022 ONSC 5268, 473 D.L.R. (4th) 609, at para. 20). Serious harm may be established through a single consideration or may arise from a combination of factors, given the holistic nature of the assessment mandated by s. 23 (C.A. reasons, at para. 140, per Brown J.A.; see also Ojeikere, at para. 63).

[72]                         The analysis is also highly individualized. It should focus on the particular circumstances of the child, rather than on a general assessment of the society to which they are sent back. As was held in Onuoha (2021), regarding a request for return of two children to Nigeria, “[r]educed opportunities, undesirable social conditions compared to conditions in Canada, and inferior economic conditions are not, by themselves, a basis for finding that the children would suffer ‘serious harm’ if returned” (para. 18; see also para. 32; M. (R.A.) v. M. (Y.Y.), 2005 BCPC 259, 48 Imm. L.R. (3d) 301). While this may be relevant on the merits of custody proceedings, judges should not, in deciding jurisdiction, compare the living conditions that each country may offer.

[73]                         The stringency of the standard of “serious harm” under s. 23 may usefully be compared with that of the Hague Convention’s grave risk of harm or intolerable situation provided for in Article 13(1)(b) of the Hague Convention (Ojeikere, at para. 53). Under Article 13(1)(b), only situations that an individual child should not be expected to tolerate meet the threshold that would be a defence to the principle of automatic return (Guide, at para. 34). The Court of Appeal for Ontario has helpfully explained that the s. 23 standard is less exacting than that of the Hague Convention (Ojeikere, at para. 58; A. (M.A.) v. E. (D.E.M.), 2020 ONCA 486, 152 O.R. (3d) 81, at para. 43). In Ojeikere, Laskin J.A. observed that the main reason for this lower standard is the difference in language between the two provisions (para. 59). The words “intolerable situation” in Article 13(1)(b) of the Hague Convention are absent from s. 23, and, “[a]s a matter of statutory interpretation, the legislature must be taken to have intended not to use these uncompromising words to qualify ‘serious harm’” (Ojeikere, at para. 59). Significantly, Laskin J.A. added that the s. 23 standard should be less stringent because signatories of the Hague Convention can rely on their reciprocal recognition that the ultimate custody decision will be based on the best interests principle (para. 60). The Court of Appeal for Ontario has also recently noted that even the Article 13(1)(b) analysis “is not meant to become an in-depth analysis of the parties’ history” (Leigh v. Rubio, 2022 ONCA 582, 75 R.F.L. (8th) 251, at para. 25). Accordingly, proceedings dealing with Article 13(1)(b) should — like those based on s. 23 of the CLRA — be resolved “quickly and efficiently” (Leigh, at para. 2).

[74]                         The “serious harm” exception is often invoked in cases where children face a risk of emotional or physical abuse. Courts have, in particular, found sufficient evidence to justify the s. 23 exception based on proof of physical and psychological harm (A. (M.A.), at para. 58; Aldush v. Alani, 2022 ONSC 1536, 74 R.F.L. (8th) 113, at para. 117). But s. 23 is not limited to those circumstances. Ojeikere provided a non‑exhaustive list of considerations that can serve as useful guidelines in assessing whether serious harm arises, although they do not encompass the wide variety of circumstances in which such a risk exists: the risk of physical harm; the risk of psychological harm; the views of the children; and the parent’s claim that they would not return to the habitual residence even if the children were required to do so (para. 64). The rules applicable to deciding custody in the foreign jurisdiction may be another relevant factor when assessing serious harm.

[75]                          Given the discretionary, individualized and fact‑specific character of the serious harm analysis, a trial judge’s findings are owed deference. I agree with Hourigan J.A. that the policy rationales supporting deference in family matters, set out in Hickey v. Hickey, [1999] 2 S.C.R. 518, and Van de Perre v. Edwards, 2001 SCC 60, [2001] 2 S.C.R. 1014, apply equally to the jurisdictional question under s. 23 (para. 38). Indeed, this Court has itself recently underscored the vital importance of appellate deference in family law matters, following the approach articulated in Hickey and Van de Perre, in Barendregt v. Grebliunas, 2022 SCC 22, at para. 103. Martin J. likewise explained, for a unanimous Court in B.J.T. v. J.D., 2022 SCC 24, at paras. 52‑53 and 55‑59, that deference on appeal is justified by an institutional difference between trial and appellate courts, where the former hear the parties directly. These justifications apply with equal force in the jurisdictional context of the s. 23 analysis because of the very character of the “serious harm” evaluation. Accordingly, appellate courts cannot set aside a trial judge’s conclusion on serious harm simply on the basis that they would have weighed the evidence differently.

(a)           Separation of Children From Their Primary Caregiver

[76]                         The prospect of separation of children from their primary caregiver, recognized in the last Ojeikere factor, raises the question of whether separation can pose a risk to the child’s psychological well‑being rising to the level of serious harm. The issue is particularly pressing in the case of young children with an established dependence on their primary caregiver parent who, as in this case, is the abductor and who is unable or unwilling to return. The majority of the Court of Appeal concluded that serious harm does not automatically arise whenever young children are separated from their primary caregiver (paras. 92‑94). Lauwers J.A., in dissent, was of the view that serious harm is established as soon as the court is convinced, on a balance of probabilities, that an “indefinite” separation between infants and their primary caregiver will occur (para. 291). Relying upon the dissenting reasons, the Mother argues before this Court that the trial judge erred in concluding that separation between children and their primary caregiver does not constitute serious harm within the meaning of s. 23 of the CLRA (A.F., at paras. 67 and 82).

[77]                         Separating an infant from their primary caregiver is a circumstance that most certainly can cause psychological harm to the child. It should never be considered lightly, given “the particular role and emotional bonding the child enjoys” with their primary caregiver (Gordon v. Goertz, [1996] 2 S.C.R. 27, at para. 121, per L’Heureux‑Dubé J.). I accept, as well, that a child’s loss of contact with a primary caregiver who cannot return to the home jurisdiction, for reasons beyond their control, can be the source of serious harm for an infant.

[78]                         But I reject the argument that such a separation, in and of itself and without regard to the individualized circumstances, will always rise to the level required under s. 23. As Hourigan J.A. wrote, this could ultimately defeat the legislative objective of discouraging child abductions of very young children (paras. 93‑94; see also Jamali v. Gillani, 2021 BCSC 2134, at para. 101 (CanLII)). In order to deter and remedy child abductions effectively, courts should be prepared, in some circumstances, to order the return of the children despite a risk of separation from their primary caregiver. Deciding otherwise could allow abducting parents, in some situations, to rely on their status as primary caregivers to circumvent the due process for custody determination and remove the children from the authority of the courts that would normally have jurisdiction. This could ultimately risk making Ontario a haven for child abductions.

[79]                         In Thomson, this Court explicitly endorsed the idea expressed in Re A. (A Minor) (Abduction), [1988] 1 F.L.R. 365 (Eng. C.A.), that “the risk has to be . . . something greater than would normally be expected on taking a child away from one parent and passing him to another” (p. 597; see also Ojeikere, at para. 56). In Bolla, Harvison Young J. (as she then was) correctly relied on this passage to conclude that the “harm” contemplated in s. 23 must be more than just the risk of harm resulting from being separated from their primary caregiver (para. 137; see also S.A.G. v. C.D.G., 2009 YKSC 21, at paras. 21 and 27 (CanLII)).

[80]                         The analysis of harm associated with separation under s. 23 must be individualized and based on evidence specific to the child involved (Ojeikere, at para. 92). The child’s age, and, where relevant, their special needs and vulnerabilities, may mitigate or aggravate the risk of harm (see, e.g., Hage v. Bryntwick, 2014 ONSC 4104, at para. 63 (CanLII)). Moreover, expert evidence accepted by the trial judge in this case shows that the relationship between a child and their known alternate caregiver and the left‑behind parent, as well as the environment to which a child will be returned, may also affect how an abducted child responds to the separation from their primary caregiver (trial reasons, at para. 309). Therefore, if a child is separated from their primary caregiver, but is nevertheless returned to their capable left‑behind parent and other known caregivers, in a safe and familiar environment, the high threshold of harm may not be met (for the Hague Convention, see Guide, at paras. 64‑65). Conversely, if the evidence demonstrates that the child would be returned to an environment where they will be left without care or that they feel unsafe with their alternate caregiver, it is very possible that the serious harm threshold will be met (see, e.g., Aldush, at para. 158).

[81]                         When assessing the severity of the harm, judges should also consider whether undertakings made by the left‑behind parent to the primary caregiver in the proceedings — also called “protective measures” — could be joined to the return order made pursuant to s. 40 of the CLRA in order to lift the obstacles to the parent’s return or to address any other aspect of the anticipated risk of harm to the child (for the Hague Convention, see Guide, at paras. 36‑37 and 64‑66; see also R. Schuz, The Hague Child Abduction Convention: A Critical Analysis (2013), at pp. 291‑92; P. R. Beaumont and P. E. McEleavy, The Hague Convention on International Child Abduction (1999), at pp. 156‑57; for non‑Hague Convention disputes in Ontario, see Bolla). These could include, for example, an undertaking from the left‑behind parent to facilitate daily contact between the abducting parent and the child or to attend to financial or administrative obstacles to the primary caregiver’s return (see, e.g., P. (J.) v. P. (T.N.), 2016 ABQB 613, 90 R.F.L. (7th) 211, at paras. 52‑60). Of course, protective measures only attenuate the risk of harm if there is satisfactory evidence that they would be respected and enforceable in the foreign jurisdiction (Schuz (2013), at p. 292). I will address the role of undertakings within s. 40 orders in greater detail below.

[82]                         As to the likelihood of separation, courts should consider all barriers to the return of the primary caregiver. In some cases, the risk of separation will be involuntary, namely when the primary caregiver faces definitive legal obstacles to their return. It may also be voluntary, in the sense that it flows from the parent’s refusal to return. In general, a parent ought not to be able to create serious harm and then rely on it through their own refusal to return (Ojeikere, at para. 91; see also Schuz (2013), at pp. 280‑81; Guide, at para. 72). Accordingly, courts should carefully scrutinize refusals to return when there is no impediment to the parent re-entering and remaining in the country of the child’s habitual residence (see, e.g., Ajayi, at para. 28). In discussing harm caused by separation from a primary caregiver who is the abductor under the Hague Convention, scholar Rhona Schuz observes:

An obvious strategy for a primary carer abductor is to state that she is not prepared to return with the child and that the consequent separation from her will cause the child psychological harm. Such claims are almost invariably rejected by most courts . . . .

 

((2013), at p. 280, relying on C. v. C. (Minor: Abduction: Rights of Custody Abroad), [1989] 2 All E.R. 465 (C.A.), at p. 471.)

This is not a matter of blaming the taking parent for their refusal to return which, as I note below, may be justified. Courts and commentators who reject allegations of harm in these circumstances are recognizing that the refusal to return, where not justified, is not in the best interests of the child. As a general rule, these best interests are served by contact with both parents, and a stable and familiar home environment. The presumption remains that abduction from their habitual residence can be the source of distress that is contrary to the child’s welfare.

[83]                         This is not to say that a primary caregiver’s refusal to return will always be taken to be unjustified. An abducting parent may have legitimate and reasonable reasons for not returning to the foreign country, such as significant obstacles to employment or risks to safety, including evidence showing that the left-behind parent is responsible for child abuse or intimate partner violence to the primary caregiver (see, e.g., Ojeikere, at para. 91; Aldush, at para. 149; B. Hale, “Taking Flight — Domestic Violence and Child Abduction” (2017), 70 Current Legal Problems 3; N. Bala and J. Chamberland, “Family Violence and Proving ‘Grave Risk’ for Cases Under the Hague Convention Article 13(b)”, Queen’s Law Research Paper No. 91 (2017), at p. 6). That said, as a general matter, an unreasonable refusal to return cannot be said to be in the child’s best interests: the law requires that parents set aside their differences, and facilitate contact between the child and their estranged partner. The mere fact of undertaking the wrongful abduction suggests an abductor has lost sight of that idea; the subsequent refusal to return, where not reasonably justified, makes further contact between the child and the left‑behind parent difficult. Again, while this is not about punishing or blaming the taking parent, it is nonetheless true that this may be a consideration to bear in mind when examining the serious harm exception in s. 23 of the CLRA (see, e.g., Onuoha (2020), at para. 128). It should be noted, however, that the trial judge in this case did not state that a possible choice not to return by the Mother would be unjustified. Instead, he considered the effects flowing from a possible separation, whatever the cause, and decided, at the end of the day, that it was in the children’s best interests to return to their home in Dubai (paras. 381 and 387‑90).

(b)          Foreign Law as a Source of Serious Harm

[84]                         The Court must also examine whether inconsistencies between family law in the foreign jurisdiction and in Ontario should factor in a serious harm analysis. Lauwers J.A., dissenting in the court below, stated his view that the UAE’s conception of the best interests of the child is incompatible with Ontario’s equal entitlement to decision‑making responsibility recognized in s. 20(1) of the CLRA (paras. 313‑17). He took issue, more specifically, with the fact that under UAE law, decision‑making responsibilities and day‑to‑day care responsibilities are automatically assigned on the basis of gender (para. 317). Before this Court, the Mother argues that the gender‑based allocation of parental responsibilities under UAE law is a pronounced departure from Ontario’s understanding of the best interests of the child and is the source of serious harm to the children here (A.F., at para. 117). The Father responds that Lauwers J.A. substituted his own findings on foreign law for those of the trial judge, and argues that deference should be shown to the trial judge’s analysis, which was reflected in his interpretation of expert evidence for which no reviewable error has been shown (R.F., at paras. 127‑28).

[85]                         International child abduction, by nature, engages two different jurisdictions, often partaking of different legal traditions for family law. In some instances, the law will be difficult to reconcile (C. Gosselain, “Child Abduction and Transfrontier Access: Bilateral Conventions and Islamic States”, in Permanent Bureau of the Conference, Preliminary Document No. 7 (2002), at p. 6). As such, the jurisdiction from which the child is removed may not share Ontario’s understanding of the law relating to custody and access and of the legal consequences flowing from the breaking up of a family unit. Even amongst parties to the Hague Convention, all of whom presumptively agree on the paramountcy of the best interests principle, the rules of family law may vary in a way that has an impact on how the best interests principle factors into the ultimate decision on custody and access on the merits (Schuz (2013), at pp. 272 and 363‑64).

[86]                         The same phenomenon may be observed in a non‑Hague Convention setting. To be sure, the announced objective of the CLRA is that the ultimate custody decision be made based on the best interests principle (s. 19(a)). This requires that a judge seized of an application for return to a non‑Hague Convention country verify if that is to be the case. But that does not mean that the judge tests whether the principle will be applied in exactly the manner in which it would be in Ontario. As Lady Hale wrote in In re J. (A Child) (Custody Rights: Jurisdiction), [2005] UKHL 40, [2006] 1 A.C. 80, a non‑Hague Convention case in which a parent requested that their child be returned to Saudi Arabia, “[i]n a world which values difference, one culture is not inevitably to be preferred to another. Indeed, we do not have any fixed concept of what will be in the best interests of the individual child. . . . Nowadays we know that there are many routes to a healthy and well adjusted adulthood. We are not so arrogant as to think that we know best” (para. 37). As such, it is appropriate for a court seized of a return application to recognize that it does not have a “monopoly in knowing what is best for children and certainly not in knowing what is best for children who have been growing up in non‑Western cultures” (R. Schuz, “The Relevance of Religious Law and Cultural Considerations in International Child Abduction Disputes” (2010), 12 J.L. & Fam. Stud. 453, at p. 477).

[87]                         As long as the ultimate question of custody is determined on the basis of the best interests of the child, the CLRA does not prevent children from being returned to jurisdictions where the law may differ in some respects from that of Ontario. I agree with Hourigan J.A. that it “is not enough to point to differences in the law and suggest that a parent may have different rights in a foreign jurisdiction vis-à-vis Ontario” (para. 79). He further explained that the serious harm test in s. 23 of the CLRA was designed to protect the safety of children and must not be interpreted so as to permit child abduction to become an approved technique for forum‑shopping (see para. 79, followed expressly on this point in Ajayi, at para. 26; see also Aldush, at para. 28; Jamali, at para. 105). Hourigan J.A. observed the danger associated with the view that, because foreign law is different from that of Ontario, it is necessarily the source of serious harm to the children: “Such a decision would send a message to parents living in the UAE that if they unilaterally come to Ontario with their children, they will not be required by the Ontario courts to send their children home” (para. 83).

[88]                         Nonetheless, there may be instances where foreign laws are so profoundly irreconcilable with Ontario law that remitting the matter to the foreign courts would constitute serious harm within the meaning of the CLRA. Drawing the line between what is acceptable and what is not is a delicate exercise. In this case, there is no denying that the expert evidence pointed to gender-based inequalities in UAE law. But as Hourigan J.A. underscored in his reasons, the proper degree of tolerance for differences between foreign and Ontario law remains defined by the threshold of serious harm (para. 79). Judges who must decide whether to take jurisdiction under s. 23 of the CLRA should be wary of their own interpretation of foreign law, in particular given the risk of treating some countries as de jure exceptions to the CLRA regime. While such a jurisdiction may theoretically exist, Canadian courts must be mindful of undercutting the fundamental premise that return will usually be in a child’s best interests (s. 19 of the CLRA). Indeed, it should be noted that the dissenting judge in the Court of Appeal was careful to say that the UAE law did not present an “iron-clad” bar to returning children there (para. 254). The proper approach recognizes that inconsistencies between local and foreign legal regimes will usually not amount to serious harm if the best interests of the child principle remains the paramount consideration in all decisions concerning children. However, if the incompatible rule automatically applies in a manner that supersedes the best interests of the child, this will be a determinative factor in the serious harm analysis, when s. 23 is read in light of s. 19(a) of the CLRA.

[89]                         In this case, the Mother argues that the law of the UAE allocates custody and guardianship on the basis of gender which, she says, is incompatible with Ontario law and means that custody on the merits will not be decided based on the children’s best interests.

[90]                         I agree that equal treatment of parents, irrespective of their gender, is a fundamental precept of family law in Ontario and is tied to the application of the best interests of the child in custody matters. The CLRA embraces this view, as s. 20(1) provides that “[e]xcept as otherwise provided in this Part, a child’s parents are equally entitled to decision-making responsibility with respect to the child”. As such, in a case where children are to be returned to a jurisdiction where their best interests would be superseded by an inflexible rule that invariably assigns custody or decision‑making powers on the basis of gender, the foreign law could be the source of serious harm and the basis for Ontario to take jurisdiction on the merits pursuant to s. 23 of the CLRA. Conversely, the risk of harm will be mitigated if the application of the gender‑based rule is not automatic but instead depends on the best interests of the child. In sum, the question for the purpose of s. 23 is whether the best interests of the child remains paramount.

[91]                         Finally, there remains the question of which standard of review should apply in appeals seeking to disturb trial judges’ findings in relation to foreign law. The Mother in this case claims that the correctness standard should apply, relying in particular on General Motors Acceptance Corp. of Canada v. Town and Country Chrysler Ltd., 2007 ONCA 904, 288 D.L.R. (4th) 74. The Father, by contrast, submits that in this case the palpable and overriding error standard is appropriate. Hapag-Lloyd AG v. Iamgold Corp., 2021 FCA 110, has sometimes been cited as authority for this proposition. He argues further that the dissenting judge in the Court of Appeal erred by substituting his view of UAE law to that of the expert opinion accepted by the trial judge.

[92]                         The manner in which this problem came before the trial judge means that this matter of the standard of review applicable to the interpretation of foreign law need not be decided as a point of general principle here. The majority judges on appeal noted the governing authorities in the province, but reasoned that deference was owed to the trial judge’s decision to prefer the evidence of the Father’s legal expert — a jurist experienced in family law before the UAE courts and fluent in Arabic — over that of the Mother’s expert who had no such comparable expertise. In point of fact, Hourigan J.A. observed that, on the crucial question of whether the best interests of the child principle would be applied during proceedings in the UAE, the experts agreed (para. 66). Relying on Larche v. Ontario (1990), 75 D.L.R. (4th) 377 (Ont. C.A.), at pp. 378‑79, Hourigan J.A. recalled the limited scope for appellate intervention in respect of a trial judge’s appreciation of expert evidence (para. 61). I agree with Hourigan J.A.’s view that the trial judge was entitled to conclude as he did and that no basis has been shown for disturbing his view. No more need be said about the standard of review of foreign law to decide this appeal.

E.            Section 40 and Return Orders

[93]                         When a court is satisfied that a child has been wrongfully removed to or is wrongfully retained in Ontario, a return order presented by the left‑behind parent is governed by s. 40.

[94]                         Three remedies are available to a judge who declines jurisdiction over custody on the merits under s. 40 (Geliedan, at para. 31). First, judges can make an interim custody order in the best interests of the child, in order to arrange for their care until a final decision is made on the merits or until it can be enforced. Judges can also stay the application on conditions, which allows them to delay the children’s return until they are satisfied that proper arrangements have been made and that the competent authorities are seized of the dispute, if necessary (see, e.g., A. (M.A.), at para. 78). Finally, judges can order the return of the child to the place they deem appropriate (Geliedan, at para. 31).

[95]                         Under the Hague Convention, judges do not have the same discretion: they will be obliged to return the children to their state of habitual residence, with or without undertakings, unless one of the exceptions apply (see Guide, at para. 19; see also Geliedan, at paras. 34 and 69; A. (M.A.), at para. 71). Despite this difference, I agree with Brown J.A. that s. 40 para. 3 return orders share the same purposes as those made under the Hague Convention (para. 127; see also Bala, at p. 308). This Court’s account of the goals of a return order in Balev is helpful here: first, to protect children against the harmful effects of wrongful abduction; second, to deter parents from abducting children in the hope that they will be able to establish links in a new country that might ultimately award them custody; and, third, to ensure the speedy adjudication of the merits in the forum of the children’s habitual residence (paras. 25‑27).

[96]                         Judges’ discretion in determining what order will best remedy an abduction in non‑Hague Convention cases is limited by the interim nature of the s. 40 powers. It follows that judges, after having determined that they should decline jurisdiction with respect to that child, cannot indefinitely delay the return of a child who is being wrongfully retained in Ontario. Had compelling reasons to retain the child in Ontario permanently been established by the evidence, it would have been relevant at the preliminary stage of determining jurisdiction. If the evidence was insufficient to establish that Ontario courts should assume jurisdiction, judges should not use their residual s. 40 powers to postpone indefinitely the child’s return to the jurisdiction best positioned to decide the case on the merits (see C.A. reasons, at para. 137, per Brown J.A.; A. (M.A.), at para. 38).

[97]                         As with any decision affecting children, judges should consider the best interests of the child in exercising their s. 40 powers (C.A. reasons, at paras. 179‑81, per Brown J.A.; M.M. v. United States of America, 2015 SCC 62, [2015] 3 S.C.R. 973, at para. 146). However, due to the interim nature of the powers, courts should not embark on a detailed analysis of the best interests factors set out in s. 24(3) of the CLRA at this stage. Brown J.A. rightly rejected the idea that consideration of the best interests of the child under s. 40 “must be performed in the same way as when determining custody (decision‑making responsibility/parenting time) under CLRA s. 24” (para. 182; see also paras. 183‑84). Section 40 orders are not custody orders on the merits.

[98]                         Nevertheless, judges must ensure the order itself properly protects the child’s interests. Incorporating undertakings from the parties within the return order may effectively facilitate a child’s return by providing an answer to the anticipated risk of harm (Bolla, at paras. 140‑45; J. Chamberland, “Domestic Violence and International Child Abduction: Some Avenues of Reflection” (2005), 10 Judges’ Newsletter on International Child Protection 70, at pp. 71‑72). Further, as in Bolla, even without a risk of “serious harm” within the meaning of s. 23, undertakings pursuant to s. 40 may be in the child’s best interests in that they effectively mitigate less consequential or short-term distress. Undertakings are well‑known and widely accepted as protective measures in international abduction cases around the world (Droit de la famille — 15751, at para. 34). I note that this Court has endorsed the use of undertakings made by the left‑behind parent to assist the return of the children in the Hague Convention setting. I agree with the view, expressed by La Forest J., that such undertakings can assist in crafting a solution that is in the best interests of the children (Thomson, at p. 599; see also Cannock v. Fleguel, 2008 ONCA 758, 303 D.L.R. (4th) 542, at paras. 26‑27). I note too that scholars have observed the prevalence and the usefulness of such undertakings, notwithstanding issues relating to their enforceability in the foreign jurisdiction, as a means of safeguarding children from the harm alluded to in Article 13(1)(b) of the Hague Convention, the rough parallel of s. 23 (J. Chamberland, “Country Report — Canada” (2005), 9 Judges’ Newletter on International Child Protection 75, at p. 79; see also M. Bailey, “Canada’s Conflicted Approach to International Child Abduction”, in B. Atkin, ed., The International Survey of Family Law (2016), 81, at p. 92). As such, courts have ordered the return of children to foreign jurisdictions, subject to undertakings given by the left‑behind parent that they would assist the abducting parent in addressing obstacles, such as obtaining immigration status that would allow them to reside in the foreign jurisdiction (Guide, at para. 68; Bailey, at p. 92; see, e.g., R.F. v. M.G., [2002] R.D.F. 785 (Que. C.A.), at para. 37; Brown v. Pulley, 2015 ONCJ 186, 60 R.F.L. (7th) 436, at para. 199).

VI.         Application to the Facts

A.           No Demonstrated Error in the Trial Judge’s Conclusion on Serious Harm

[99]                         I would reject the Mother’s submission that the trial judge committed a palpable and overriding error when he concluded that the children would not suffer serious harm if they were returned to Dubai. While it is unquestionable that s. 23 calls for a holistic and individualized assessment of serious harm (Ojeikere, at para. 63; C.A. reasons, at para. 140, per Brown J.A.), there are two factors that are of particular importance in this case: separation from the primary caregiver and incompatibilities between UAE and Ontario law. Accordingly, I will address these two factors before turning to whether the trial judge erred in his holistic assessment of whether the children would suffer serious harm.

(1)          Separation From Primary Caregiver

[100]                     I agree with the view of the majority of the Court of Appeal that no reviewable error has been shown in the trial judge’s conclusion that the anticipated impact of separation does not meet the s. 23 threshold. Respectfully, I disagree with the dissenting judge that the trial judge erred by finding that the specific impact of separation on Z. and E., was “unknown” (para. 287).

[101]                     When the trial judge’s reasons are read as a whole, it is clear that he properly drew individualized conclusions as to whether the children in this case would risk suffering “serious harm” within the meaning of s. 23 as a result of their separation from their mother. Considering a trial judgment in its totality is especially important where the judge weighed the evidence and made factual findings throughout their judgment, rather than presenting all of their analysis on an issue in one section. This is what happened in the trial judge’s decision in this case. For example, the judge found that both the Mother and the Father are “loving, caring parents” (para. 15), and that the children’s long‑term nanny has been very involved in their care in Dubai (para. 291(ii)). Both of these are examples of considerations that form part of the broader factual context relevant to the ultimate finding that there was no serious harm in this case, a finding that deserves deference here (Van de Perre, at para. 15).

[102]                     The broader factual context is also the reason why the Mother is wrong to claim that the trial judge concluded that the potential impact of separation on Z. and E. was “unknown” (A.F., at para. 151). Before coming to the section where he stated his conclusions on the serious harm analysis, the trial judge reviewed the expert evidence, and accepted that, generally, when infants are separated from their primary caregiver, they can face emotional and psychological distress (para. 294(iii)). Accordingly, reading the trial judgment as a whole suggests that he did not say that the impact on the children in this specific case was “unknown” (para. 294(iv)) because he mistakenly declined to decide the issue or because he failed to recognize that they would suffer psychologically. Instead, the comment relates to his finding that the expert evidence was unreliable on the specific impact of the separation on Z. and E. (paras. 313‑15). As Hourigan J.A. pointed out, the expert herself conceded on the voir dire that she could not testify about the psychological impact of separation on Z. and E. with “any degree of certainty” (trial reasons, at para. 234; C.A. reasons, at para. 90). The trial judge nevertheless conducted his own individualized assessment of serious harm. He observed that infants “can face serious negative effects from being removed from their primary caregiver” (para. 305) and extended this finding to conclude that Z. and E. could be at risk of emotional and psychological harm if they were returned to Dubai without their mother (para. 366(ii)). To the extent that it could be interpreted as his final conclusion on the matter, his statement that the impact on Z. and E. was unknown is at most an unfortunate formulation. But this did not end his analysis. The trial judge was aware that he did not need to accept particularized evidence about the children in this case to conclude that they would suffer from being separated from their mother.

[103]                     Despite his finding that Z. and E. were facing a risk of emotional and psychological distress, the trial judge concluded that this contemplated risk did not rise to the required level of “serious harm” within the meaning of s. 23 (para. 370). He was entitled to distinguish between adverse emotional effects and the high bar of serious harm required for assuming jurisdiction. Put another way, he was entitled to conclude that, while there was undoubtedly harm, “serious harm”, within the meaning of s. 23, was nonetheless not present. His conclusions are owed deference and cannot be disturbed absent an error in law or a material error in the appreciation of the facts (Van de Perre, at para. 13). I see no reason to interfere with the trial judge’s conclusion that the adverse emotional impact of separation on the children, weighed by its likelihood and severity, did not amount to “serious harm” under s. 23 of the CLRA.

[104]                     The record indicates that this conclusion was open to the trial judge. While Z. and E. are young, the trial judge observed that they do not have special needs or vulnerabilities (para. 14) which would aggravate the seriousness of the harm flowing from their separation from their mother. Importantly, the trial judge concluded that the Father, like the Mother, is a good, loving and caring parent. There is not “a hint of evidence” of the children being mistreated or abused by the left‑behind parent (paras. 15 and 480). According to the expert evidence accepted by the trial judge, the quality of the alternate care that Z. and E. would receive diminishes the seriousness of the harm flowing from the separation (para. 309).

[105]                     The trial judge also considered the Father’s parenting plan. He accepted that if the Mother does not return to Dubai with her children, the live‑in nanny — who has played a significant role in the upbringing of Z. and E. (paras. 261‑62 and 291(ii)) — and other relatives will be in Dubai to assist the Father, himself a capable parent, with childcare (para. 52). It may be recalled that the Mother disputed the role Mary played in respect of the children at trial; the trial judge rejected the Mother’s evidence on this point as not credible (para. 262). Further, the children would return to a safe and familiar living environment, and Z. would also return to a known school. These were each factors that an expert before the trial judge identified as potentially helping to protect against some of the harms caused by separation from a parent (para. 236). Taken together, these factors could fairly be understood to mitigate the severity of the harm that might be caused to Z. and E. as a result of a possible eventual separation from their mother. While, of course, they could never entirely make up for the absence of the primary caregiver, they are relevant to the overall assessment of whether these children would suffer serious harm upon separation. The Mother has shown no reversible error in the trial judge finding that there was no risk of serious harm within the meaning of s. 23 of the CLRA. This is especially so in view of the trial judge’s plain finding, made in respect of s. 40 of the CLRA, that the return of the children would be in their best interests (paras. 381 and 387‑90).

[106]                     I turn next to the trial judge’s assessment of the likelihood of separation. The Mother repeatedly asserted that she would not return to Dubai, even if her children did. The trial judge found that she was not credible on this point and decided to place very little weight on it in his “serious harm” analysis (para. 368). His conclusion as to the Mother’s credibility is also owed deference. It should be recalled that the trial judge was not assessing the Mother’s credibility solely to determine the likelihood of separation and consequent harm. The Mother’s credibility was important for other factual issues. For instance, whether the Father had ever been physically aggressive (paras. 272‑74) and the extent of the assistance Mary provided in the care of the children (paras. 261‑62); on both these points the Mother’s evidence was rejected. However, the trial judge was careful to make no definitive finding — whether based on her credibility or any other evidence — that the Mother would or would not return. The trial judge did express his hope that the Mother would go back (para. 380). But this hope did not blind him to the possibility that she might not do so. For instance, the trial judge explicitly considered what would happen to the children if “she chooses . . . to not return to Dubai” (para. 381). Likewise, the trial judge said that he was “not sure” the Mother would remain behind in Ontario — but immediately thereafter turned to consider what his assessment would be if he did believe her evidence about not returning to Dubai (para. 368).

[107]                     At each stage, the trial judge remained alive to the Mother’s desire not to return, and to the very real possibility that she might remain behind in Ontario even if the children were ordered to return. This possibility was a foundational premise of his assessment of the likelihood of serious harm. The trial judge wrote that he was “not satisfied on balance that Z. and E. would suffer serious harm if they are removed from Ontario and returned to Dubai: s. 23” (para. 389). He was of that view whether or not the Mother returned to Dubai with them, as he was convinced that their return would be in their best interests (para. 387). The trial judge returned to this conclusion in the penultimate paragraph of his reasons, saying that “it is time for these children to go home. No harm will result to them. Their father is there. Their nanny is there. The [school Z. attended] is there, by all accounts a first-class facility. Their lives are there. Their mother will continue to be their primary caregiver, if she sees fit to return. If she does not, then it remains the case that the children should go home” (para. 481). The Mother has shown no reversible error in this assessment. To be clear, the judge’s conclusion was not a value judgment on whether or not the Mother should return: like in Ojeikere, “[t]here may be cases where a parent’s refusal to accompany the children back to the country of habitual residence could give rise to a serious risk of harm to the children. This case is not one of them” (para. 92).

[108]                     Mindful that she would be a divorced non‑national, the trial judge also considered the legal barriers the Mother would face if she sought residency in the UAE. On the basis of evidence given by Ms. Hamade, who was apprised of the Father’s settlement offer, the trial judge accepted that there were several options available to the Mother should she decide to return to Dubai with her children (paras. 194 and 293‑94). Importantly, the trial judge considered whether any “protective measures” proposed by the Father in his “with prejudice” letter could diminish the likelihood of separation, and ultimately, alleviate some of the distress suffered by Z. and E. He underscored that the Father offered, in a settlement agreement that could have been incorporated within a s. 40 para. 3 order, to secure independent residency status for the Mother (paras. 49 and 472). On the basis of expert evidence, the trial judge unreservedly concluded that such an agreement, if accepted by the Mother, would be able to be incorporated into an enforceable, valid order by Dubai courts (para. 294(ii)).

[109]                     Lauwers J.A., without directly challenging these findings, suggested that the trial judge failed to consider the fact that the options available to the Mother to secure her UAE residency were not “assured” (para. 292). For the dissenting judge, this uncertain residency status meant that the children will most likely be separated from the Mother and serious harm will necessarily follow (para. 291). I respectfully disagree. First, it appears to assume that in all cases separating infants from their primary caregiver automatically constitutes serious harm. Second, it suggests that the trial judge did not consider the extent of the harm that the children would suffer if the Mother did not return with them and the separation materialized. The trial judge concluded however that the children would not suffer serious harm if returned to Dubai, with or without the Mother and further noted it would be in the children’s best interests to be returned there, even without her (paras. 387 and 389). Third, it ignores the fact that the trial judge was, as Hourigan J.A. pointed out (para. 63), aware of concerns regarding the precarious residency status of the Mother and of the undertakings the Father had made to dispel those concerns. The trial judge also offered the Mother an opportunity to make further submissions on the possibility to include the Father’s settlement proposal in the return order intended to mitigate the harm caused to the children (para. 472). The Mother did not avail herself of that opportunity (C.A. reasons, at paras. 62‑63). As I note below, the trial judge understood the performance of the undertakings was not guaranteed but fairly accepted that they could be incorporated into an enforceable court order in Dubai (para. 301).

[110]                     But even taking the Mother’s case at its highest, recognizing that she “cannot and will not return to the UAE” and that separation with her children will definitely occur, I cannot give effect to her argument that the trial judge failed to acknowledge the serious harm that a separation would cause to the children (A.F., at paras. 68 and 73 (emphasis deleted)). The trial judge heard the parties and the expert evidence, and he came to the conclusion that the s. 23 threshold was not met. He concluded that it would be in the children’s best interests to return to Dubai, even if the Mother did not follow them (para. 387). The Mother did not demonstrate that this conclusion was unsupported by the evidence or otherwise reflects a palpable and overriding error.

(2)          Foreign Law and the “Best Interests” Standard in the UAE

[111]                     The Mother also argues that the trial judge erred in concluding that parenting decisions in the UAE are made according to the best interests of the children. She says the fact that foreign law refers to the “best interests of the children” (A.F., at para. 108) is insufficient: what matters is whether the substance of that standard conforms to Canadian law. In her view, the trial judge failed to consider significant differences between Canadian and UAE law in this regard (paras. 108‑16).

[112]                     The crux of this argument is that pursuant to the relevant UAE legislation — cited here as Federal Law No. 28 of 2005 on Personal Status — parental responsibilities are presumptively assigned on the basis of gender. Mothers are granted “custody” of the children, while fathers are granted “guardianship”. The two roles involve different responsibilities and decision‑making powers. While the custodian’s role is to ensure “day‑to‑day” care, the guardian is responsible for important financial decisions and other major life decisions, including the children’s education and religion. A mother’s custodial rights can be terminated if she remarries, and when a male child reaches the age of 11, or when a female child reaches the age of 13.

[113]                     The Mother argues that this gender‑based allocation of parental responsibilities is radically incompatible with gender equality in Ontario family law (A.F., at paras. 101 and 117). Lauwers J.A. described it as a pronounced departure from Ontario’s understanding of the best interests of the child (para. 316). While recognizing that a foreign law need not replicate Ontario’s rules, he wrote that the fact that it might fall short of domestic standards must be a relevant consideration for determining best interests of the child (para. 305).

[114]                     I agree with Lauwers J.A. that the division of parental responsibilities on the basis of gender is inconsistent with the gender equality upon which the allocation and exercise of custody and access rights rests in Ontario law. As he noted, a child’s parents are equally entitled to decision‑making responsibility, and determinations of custody are made on the sole basis of the best interests of the children (ss. 20(1) and 24(1) of the CLRA). The inconsistency with Ontario law with which the Mother takes issue cannot be discounted as minor or purely technical: it touches upon a fundamental principle of Ontario family law. This significant discrepancy required the trial judge to determine whether the best interests of the child principle would nevertheless prevail under UAE law should he order the return of Z. and E. for the ultimate custody decision, an objective announced in s. 19(a) of the CLRA.

[115]                     The trial judge did conclude that the best interests of Z. and E. would be the paramount consideration in determining custody in a court in Dubai.

[116]                     In reaching his conclusion that the children would not suffer serious harm if Dubai courts determined issues of custody on the merits, the trial judge was aware that aspects of UAE law are said to conflict with Ontario’s conception of the best interests of children. He relied on the testimony of both experts to examine how those rules are applied by UAE courts in practice (paras. 187‑216). He concluded, on the strength of that evidence, that the provisions mandating the allocation of parental responsibilities on the basis of gender are not automatic or imperative, but are rather subject to the discretion of judges who ultimately decide custody and access on the basis of the best interests of the child (paras. 294‑304).

[117]                     Specifically, he relied on the evidence of Ms. Hamade, the Father’s expert, who testified that the rules according to which a mother’s custodial rights can be terminated if she remarries, or when a male child reaches the age of 11 or a female child the age of 13, do not automatically apply. The best interests of the child principle always takes precedence. As Ms. Hamade put it, “[t]he court will always look at the best interest of the children in deciding whether they should be removed from the mother, even if the mother remarries. . . . [T]he best interest of the children would always be the governing criteria for the [c]ourt to decide whether the children should stay or be removed” (A.R., vol. IV, at pp. 224‑25; trial reasons, at para. 185).

[118]                     Moreover, the trial judge underscored the importance of the cross‑examination of the Mother’s expert to his ultimate conclusion on this point. The trial judge explained how, in cross‑examination, Ms. Schildgen conceded that any decision regarding the children by a UAE court would be based on their best interests and that an abundant jurisprudence from the UAE “make[s] it patently and obviously clear that it is the best interests of the child that will determine all decisions of the [c]ourt”, including in the event that the mother remarries (A.R., vol. VII, at pp. 43‑46; trial reasons, at paras. 215‑16 and 303‑4).

[119]                     The trial judge relied on the testimony of both experts to support two distinct findings with respect to UAE law that were central to his conclusion that its application did not constitute serious harm: first, that the best interests of the children would be the paramount consideration in determinations of custody, access and guardianship; and second, that an agreement between the parties could be incorporated into a valid and enforceable court order (paras. 294 and 298‑99). Indeed, the “Divorce Settlement Agreement through Dubai Courts” — a document that was before the trial judge — included terms that, if agreed to by the Mother, would provide her with custody of the children until they reach the age of 18, even in the event that she remarries (A.R., vol. XI, at pp. 310‑11, cls. 3/1 and 3/2).

[120]                     The crucial point here is that the trial judge did not reach his conclusion on serious harm because he failed to consider the impact of the provisions of the UAE law with which the Mother takes issue. Not only did he consider expert evidence on how the courts in Dubai apply these provisions in practice, but he turned his mind to the practical consequences the impugned provisions would have on Z. and E., considering the circumstances of their family. Hourigan J.A. was right to underscore that the trial judge found “the parties agreed on significant issues like schooling and religious instruction” (C.A. reasons, at para. 81; trial reasons, at para. 378). And if the parties eventually disagreed, Dubai courts would decide according to Z. and E.’s best interests. On this last point, the trial judge also considered that the best interests of the child principle may carry a different meaning under UAE law and acknowledged that Dubai courts might not look at the same factors as those provided for by s. 24 (para. 454). He rejected the Mother’s main argument that Dubai courts would not “really” employ a “best interests” analysis — a claim in respect of which even her own expert disagreed (para. 367).

[121]                     These findings are owed deference. Judges interpret foreign law on the basis of expert evidence, their findings should not be interfered with in the absence of an overriding and palpable error. The Mother is mistaken to invite us to interfere with the trial judge’s reading of the expert evidence and to conclude, based on our own interpretation of UAE law, that a parenting determination in Dubai would not be based on the children’s best interests. The Mother’s position appears to be that the absence of a best interests regime is endemic to UAE law — a position that even the dissenting judge rejected (C.A. reasons, at para. 254).

[122]                     As a result, the Mother has offered no principled basis to revisit the trial judge’s conclusion that, despite the diverging conceptions of family law in the UAE and in Ontario, the children would not be exposed to serious harm if returned to Dubai.

(3)          Conclusions on Serious Harm

[123]                     Assessing the application of the s. 23 regime in this case, I am satisfied that the trial judge properly considered the relevant constellation of factors in this case, weighed them in light of the credibility of the witnesses, and based his conclusion on an individualized child‑centered analysis. As the trial judge recalled, both parents are capable, loving parents; he said, notwithstanding some of the Mother’s allegations, that there is not a hint of evidence that she or the children would be subject to abuse if they return. But the trial judge went further, weighing the other relevant factors. For instance, the dissenting judge said the trial judge failed to assess the harm of an involuntary separation. I respectfully disagree: the trial judge considered the Mother’s residency status in light of the undertakings; he considered the expert evidence on the options available to the Mother; and he decided that the best interests of the children would be the paramount consideration under UAE law. His conclusion was not based on a misplaced view that the Mother was certain or obliged to return; instead, he assessed the actual harm that would flow from the return of the children, whether or not the Mother went with them. His finding that the serious harm threshold was not met is owed deference, and the Mother has not persuaded me that he erred. Another judge might have decided serious harm differently, but that is no basis for disturbing the conclusion of the trial judge given the applicable standard of review.

B.            The Father’s Undertakings Are Included in the Return Order

[124]                     The Mother submits that the trial judge did not properly exercise his s. 40 powers. Before ordering the return of the children, he should have considered their best interests comprehensively, particularly in light of the Mother’s precarious residency status in the UAE and of the impact of the return order on the determination of custody in the UAE (A.F., at paras. 145‑46). She relies on Lauwers J.A.’s comment that the decision to make a return order under s. 40 para. 3 of the CLRA requires a distinct “best interests of the child analysis” (paras. 335 and 339).

[125]                     I respectfully disagree with the Mother.

[126]                     As noted, when judges decline jurisdiction, they are not entitled to rely upon their s. 40 powers to make determinations on the substance of the dispute. The Mother cannot renew her argument to obtain, at this stage, the result she was seeking in asking the court to assume jurisdiction. Section 40 powers should not be used to frustrate the return of the children to their habitual residence when foreign authorities are judged to be the appropriate adjudicators following a wrongful removal to, or wrongful retention in, Ontario.

[127]                     Furthermore, it is apparent that the trial judge did consider the best interests of the children when he made his decision. In his opinion, it was in the best interests of Z. and E. to be returned to Dubai (paras. 381, 387 and 481). As Brown J.A. observed, the trial judge made several specific findings of fact regarding the interests of the children, “about the relationship between the children and both parents and the potential impact of a return order on the children; the adequacy of evidence of the best interests of the children available in Ontario; the potential risk of harm to the children if returned to Dubai; that the settlement proposed by the father, if accepted by the mother, would be incorporated in a Dubai judgment; and that the best interests of the children would be the paramount consideration in determining custody in a court in Dubai” (C.A. reasons, at para. 189; trial reasons, at paras. 294, 347 and 366).

[128]                     Having declined to exercise jurisdiction and having concluded that removal from Ontario would not expose the children to serious harm, the trial judge was entitled, by authority of s. 40 of the CLRA, to order their return to Dubai. The Mother raises no valid grounds for intervention on this point. It also bears recalling, as did the majority in appeal, that the return order is not a custody order and does not result in the loss of any legal rights by the Mother. This appeal decides jurisdiction, not custody. That will be decided by the competent court in the UAE.

[129]                     What remains to be decided is whether this Court should include some of the undertakings proposed by the Father as part of his settlement offer in our disposition of this appeal.

[130]                     In his submissions before our Court, the Father restated his argument that the “with prejudice” undertakings he had made at trial would adequately remedy the precariousness of the Mother’s status in Dubai if she returned with the children and, notwithstanding the expiry of any period of eligibility for residency for the Mother, he renewed those undertakings. He noted that the trial judge relied on the expert Ms. Hamade’s explanation that the Mother had various options available to her upon return, including the promised sponsorship by the Father and the purchase, by him for her, of property in Dubai (trial reasons, at paras. 194‑96). In her written argument before this Court, the Mother acknowledges, apparently for the first time in this litigation, the idea that undertakings could have answered the uncertainties surrounding her return to Dubai. She suggests that had they not been insufficient, the measures could have displaced concern that the children would face harm on separation (A.F., at paras. 9, 148 and 153). The Mother also includes, in her condensed book filed in this Court, the letter counsel to the Father sent her, which was referred to by the trial judge in his reasons (trial reasons, at paras. 49 and 472), in which the Father restates that his undertakings could be part of an eventual s. 40 return order.

[131]                     I consider the Mother’s acknowledgment of the potential usefulness of undertakings and the Father’s renewal of his commitment to take measures to remedy her precarious situation on return as an expression of willingness, on both sides, that undertakings be part of the s. 40 return order.

[132]                     The trial judge, in the exercise of his discretion under s. 40, considered the undertakings to be adequate. I see no reason to disturb this conclusion but, respectfully, they should have been made explicit in the order. I note, however, that the trial judge called for submissions by counsel on the undertakings made by the Father in the settlement proposal (para. 472) and, as noted by Hourigan J.A. on appeal, the Mother did not respond (paras. 62‑63). If the trial judge had had the benefit of the Mother’s views on the undertakings, he would likely have explicitly included them in his order. I note his reliance on Ms. Hamade’s expert testimony, who took them into account, and his lengthy quotation of the undertakings at para. 49 of his reasons.

[133]                     Given the paramountcy of the best interests of the child principle to a s. 40 order, the refusal of the Mother to provide submissions on this point should not have prevented the trial judge from including the undertakings in his order. They provide a source of comfort to the Mother who is concerned about her status if she does decide to return and mitigate the risks to the children associated with separation from her. But to be plain, they were not the basis for the trial judge’s conclusion that there was no “serious harm” here, a conclusion that he came to independently of the undertakings. Acknowledging the role of the undertakings is in the best interests of the children. I would thus make plain that, in my view, the return order recognized the undertakings as a means of attenuating the Mother’s precarious status upon her return to Dubai and, as a result, of limiting the adverse effects of such a return on the children, should she choose that route. This appears to be the most favourable circumstance for the children, who would benefit from both their parents’ continued presence and involvement in their lives.

[134]                     Problems associated with the enforceability of undertakings by foreign courts are well known. But, in this case, the trial judge unreservedly concluded, on the basis of the expert evidence, that the settlement offer proposed by the Father — if agreed to by the Mother — could be incorporated into a valid court order in Dubai, and would be enforceable (para. 294(ii)). In deciding that the undertakings were a viable commitment from the Father, the trial judge plainly had no guarantee that the Father would not later resile from that commitment, as Lauwers J.A. emphasized in his dissenting reasons (paras. 301‑2). The trial judge did understand that the commitments were not guaranteed but I respectfully disagree that he put blind faith in the Father on the basis of his general finding of credibility. The undertakings might have been made a condition for the return but, as commentators note, this is not always the practice (Schuz (2013), at pp. 291‑92). In my view, what is required is that the judge who hears the parties is satisfied that the undertakings given are adequate (see Schuz (2013), at p. 290). This assessment is discretionary and must be made in light of the parties’ particular circumstances. I agree with Hourigan J.A. that nothing in the record indicated that the Father would not respect his undertakings (paras. 71‑72).

[135]                     Moreover, I share the view of Chamberland J.A., writing in a Hague Convention case, that a left‑behind parent has no advantage in acting in bad faith and reneging from their commitments in the foreign jurisdiction: this is a fact that the foreign authorities deciding custody would consider on the merits, and those same authorities [translation] “would certainly hold against him in a context where the best interests of the children will be at the centre of the analysis” (Droit de la famille — 15751, at para. 36). In other words, when the ultimate custody and access determinations are made in a UAE court, the Father will have every interest in showing that he respected undertakings made in the best interests of the children. It was not unreasonable for the trial judge to rely on the undertakings here.

[136]                     In this case, I find it useful, then, to make explicit the undertakings the trial judge himself considered to be of assistance to remedy the Mother’s precarious status. I recall that when undertakings are made under s. 40, the absence of “serious harm” has already been decided upon, as s. 40 para. 1 makes plain. These undertakings diminish one source of the risk of involuntary separation between the Mother, as primary caregiver, and the children upon return to Dubai. As evidenced in the Father’s letter of October 30, 2020, and recited by the trial judge at para. 49 of his reasons, I acknowledge that the Father is bound by the following undertakings:

N. undertakes as follows:

 

1. Prior to F.’s returning to Dubai to live with the children, N. shall vacate the home that the parties’ [sic] and the children lived in prior to F. leaving Dubai. The contents of that property will remain at the property, subject to N. removing his personal items (clothing etc.). N. will tend to payment of all expenses for the property. F. can live with the children at the property until the property below (the “new property”) is purchased and is available for occupation by F.

 

2. Mary will continued [sic] to be employed to assist with the care of the children when they are at N.’s residence. Another nanny will be hired to assist with the care of the children at F.’s home.

 

3. Within 90 days of F. returning to Dubai, N. will purchase of [sic] a home in Dubai, with the value of up to 1,000,000 Emirati Dirham (AED). He will consult with F. with respect to the selection of that property and she can decide what property will be purchased. If she wishes to purchase a property that is worth in excess of AED 1,000,000, she can either finance that amount or make that contribution herself. If that property has been purchased before F. returns to Dubai with the children, she will move to and occupy that property, and N. will continue to occupy the current home.

 

4. Title to the new property will be taken in F.’s name, which will ensure her independent residency status as registered owner of land, but there would be a separate trust document that would confirm that she holds title in trust for the children, with the children having equal beneficial interest in the property. The children would be entitled to realize their beneficial interest upon the youngest of the children attaining the age of 18 years, or such other time as the parties mutually agree. If F. makes a financial contribution to the purchase of the property or has arranged financing of it then that will entitle her to a direct interest in the property in proportion to that contribution.

 

5. The arrangements for the parenting/custody/access of the children will be as set out in the draft agreement attached as Exhibit A to N.’s affidavit sworn September 28, 2020 except that where terms of paragraphs 1-4 hereof conflict with the terms of the draft Agreement, paragraphs 1-4 hereof shall apply.

 

6. Financial provision for the children and F. will also be a [sic] set out in the Exhibit A to N.’s affidavit sworn September 28, 2020, except that where terms of paragraphs 1-4 hereof, conflict with the terms of the draft Agreement, paragraphs 1-4 hereof shall apply.

 

7. In the alternative to paragraphs 5 and 6, if F. wishes to pursue her rights entitlements and remedies in the court in UAE, with respect to custody / access / parenting / support issues / property, she may do so.

 

8. The terms herein will be made into a consent Order/Judgement issued by the [c]ourt in the UAE.

[137]                     I would dismiss the appeal, with costs. I would not disturb the costs orders in the courts below.

The reasons of Karakatsanis, Brown, Martin and Jamal JJ. were delivered by

 

                   Jamal J. —

I.               Introduction

[138]                     I have had the benefit of the reasons of my colleague Justice Kasirer. I agree with my colleague’s discussion of the background facts, decisions below, statement of issues, and applicable legal principles, but I part company with him in applying the law to this case.

[139]                     The main issue in this appeal is whether the trial judge erred in concluding that the Ontario courts lack jurisdiction to determine the merits of a custody dispute involving Z. and E., the parties’ two infant children aged four and one at the time of trial (now aged six and three). The children and the mother are Canadian citizens, and the father is a Pakistani citizen living in Dubai in the United Arab Emirates (“UAE”) on an employment visa. The jurisdictional issue turns on whether the mother established on the balance of probabilities that the children “would . . . suffer serious harm” under s. 23 of the Children’s Law Reform Act, R.S.O. 1990, c. C.12 (“CLRA”), if they are removed from Ontario and returned to the father in Dubai without her.

[140]                     The CLRA aims to discourage child abduction and provides for the return of abducted children (see ss. 19 and 40). At the same time, by enacting s. 23 of the CLRA, the Ontario legislature recognized that, in some cases, the objective of discouraging abduction must yield to the paramount objective of preventing serious harm to children.

[141]                     In this case, the trial judge found that the mother did not establish serious harm and made a return order under s. 40 of the CLRA (2020 ONSC 7789, 475 C.R.R. (2d) 1). A majority of the Court of Appeal for Ontario dismissed the appeal (2021 ONCA 614, 158 O.R. (3d) 481).

[142]                     I agree with my colleague that to determine the issue of serious harm, the trial judge had to assess the likelihood and severity of the anticipated harm considering all relevant factors (Kasirer J.’s reasons, at para. 71; Ojeikere v. Ojeikere, 2018 ONCA 372, 140 O.R. (3d) 561, at para. 62, per Laskin J.A.). In my view, the trial judge made material errors in assessing both likelihood and severity.

[143]                     The trial judge misapprehended the evidence relating to the likelihood that the children would suffer harm if they are separated from their mother and returned to the father. This likelihood turned on the mother’s claim that she will not return to Dubai. The trial judge declared that he was “not sure” that he believed the mother’s claim of non-return and assigned “very little weight” (para. 368) to it because of inconsistencies in her account of tangential and largely irrelevant matters. In relying only on these inconsistencies, the trial judge ignored several crucial relevant considerations supporting the mother’s claim, thereby tainting his conclusion on the likelihood of the anticipated harm.

[144]                     I reject any suggestion that the mother has “self-engineered” her claim of serious harm by steadfastly refusing to return to Dubai. As my colleague acknowledges, a parent can legitimately refuse to return to the place of habitual residence for a “substantial reason” (Kasirer J.’s reasons, at paras. 82-83 and Ojeikere, at para. 91). The mother provided reasonable and legitimate reasons for refusing to return to Dubai. Her precarious residency status in Dubai, her bases for refusing the father’s “with prejudice” settlement offer purporting to provide her with benefits if she returns, and her legitimate concerns about living under the laws of the UAE as a woman cumulatively rebut any suggestion of self-engineered harm.

[145]                     The trial judge also misapprehended the harm to the children. He did not give effect to his own factual findings about the children, including as to the parents’ respective roles in caregiving, in light of the jurisprudence and the expert evidence showing that children generally suffer harm when separated from their primary caregiver.

[146]                     I disagree with my colleague’s suggestion that the trial judge determined that the children would not suffer serious harm by relying on the mitigating effects of alternate caregivers (paras. 104-5 and 107). The proposed alternate caregivers are the children’s former nanny in Dubai and two of the father’s relatives currently living in Pakistan and the United States, one of whom has never met the children. The trial judge did not address these factors in his s. 23 analysis, and in any event, they do not adequately mitigate the obvious harm that the children would suffer if they are separated from their mother.

[147]                     The trial judge’s material errors displace the appellate deference generally afforded to discretionary determinations and demand appellate intervention. In my view, the mother met her burden of establishing serious harm. I would allow the appeal.

II.            Discussion

[148]                     The trial judge described the s. 23 issue of whether the children would suffer serious harm if returned to Dubai as “the most legitimately debatable point at trial” (para. 364). Yet his 482-paragraph trial decision addressed this point in just 8 paragraphs (paras. 363-70): 2 introductory paragraphs, 1 paragraph quoting extensively from the leading judgment in Ojeikere, and 5 paragraphs containing largely conclusory analysis or findings.^[2]

[149]                     I agree with my colleague that the trial judge weighed the evidence and made factual findings throughout his reasons (Kasirer J.’s reasons, at para. 101). But apart from paras. 363-70, none of the paragraphs of the trial judge’s voluminous reasons cited by my colleague analyzed s. 23 of the CLRA in relation to the specific circumstances of these children and their parents. Instead, the trial judge’s entire analysis of s. 23 is contained in eight paragraphs under the heading “Section 23”.

[150]                     I also agree with my colleague that a court’s s. 23 determination is discretionary and generally attracts appellate deference (paras. 11 and 75). I accept that appellate courts should not intervene simply because they would have weighed the likelihood or severity of harm differently (para. 75). But appellate deference is not without limit. An appeal court may intervene if there has been “a material error, a serious misapprehension of the evidence, or an error in law” (B.J.T. v. J.D., 2022 SCC 24, at para. 52 and Van de Perre v. Edwards, 2001 SCC 60, [2001] 2 S.C.R. 1014, at para. 11, citing Hickey v. Hickey, [1999] 2 S.C.R. 518, at para. 12). As this Court stated in Van de Perre, at para. 15, “[i]f there is an indication that the trial judge did not consider relevant factors or evidence, this might indicate that [they] did not properly weigh all of the factors. In such a case, an appellate court may review the evidence proffered at trial to determine if the trial judge ignored or misdirected himself with respect to relevant evidence”.

[151]                     As I will explain, the trial judge seriously misapprehended the evidence in evaluating both the likelihood and severity of the harm. Appellate intervention is therefore warranted.

A.           The Trial Judge Erred in Assessing the Likelihood of Harm

[152]                     The trial judge erred in assessing the likelihood that the children would be harmed if removed from Ontario by ignoring relevant evidence as to the mother’s claim that she would not return to Dubai. Because the mother has legitimate reasons for refusing to return to Dubai, she has not “self-engineered” her claim of serious harm to the children. She is therefore not prohibited from relying on harm arising from her legitimate decision not to return to Dubai.

(1)          The Trial Judge Ignored Evidence Relevant to the Mother’s Claim That She Would Not Return to Dubai

[153]                     In his s. 23 analysis, the trial judge had to consider the likelihood that the two infant children would suffer harm if returned to Dubai (Ojeikere, at para. 62). In this case, that turned on the mother’s claim that she would not return.

[154]                     The only evidence in the record on this point was from the mother. She testified that she would not return:

. . . I will not go back. My -- my -- my children, first of all and before anything else, the -- the reason for the separation was because they deserve a happy mom and deserve a mom, a parent who is happy, and so I don’t see the possibility of that happening if I am -- if I am to go back.

 

. . .

 

. . . I see myself being a better parent if I’m actually in Canada and getting access.

 

(A.R., vol. VI, at pp. 245-46)

[155]                     The trial judge placed “very little weight” on the mother’s claim (para. 368). His reason was, in his words:

I am not sure that I believe [the mother], for the reasons outlined earlier in this Judgment. [para. 368]

[156]                     The “reasons outlined earlier” were the trial judge’s reasons for concluding that the mother was not a credible witness (para. 256). In assessing the mother’s credibility, the trial judge did not identify any issues in her evidence that she would not return to Dubai. Instead, he relied on the mother’s unwillingness to correct or clarify certain other evidence, as well as tangential and largely irrelevant inconsistencies in her evidence about, for example, whether the nanny Mary helped “a lot” with the children; whether the father’s pornography consumption involved girls ages 14 to 16 or rather ages 16 to 19; and whether the mother was resident in Milton, Ontario, since 2005 (paras. 255-87).

[157]                     In assigning little weight to the mother’s claim based on these inconsistencies, the trial judge ignored relevant evidence as to why she would not return to Dubai. Such an omission is a material error because the trial judge’s reasons give rise to the reasoned belief that he must have forgotten, ignored, or misconceived the evidence in a way that affected his conclusion (see Van de Perre, at para. 15; Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235, at para. 39). The trial judge was required to turn his mind to the relevant factors that go to the believability of the mother’s claim (see R. v. G.F., 2021 SCC 20, at para. 82, per Karakatsanis J.). Although a trial judge’s credibility findings “deserve particular deference” (G.F., at para. 81), a credibility finding can be set aside on appeal when the finding is “arbitrary in that it is based on an irrelevant consideration” (Waxman v. Waxman (2004), 186 O.A.C. 201, at para. 364). Such is the case here.

[158]                     Here, the trial judge addressed none of the following factors that were relevant as to whether the mother’s claim that she would not return to Dubai should be believed:

(a)     First, the mother expressed concerns about living as a woman in the UAE. She noted that in the UAE, the police do not take reports of domestic violence seriously (para. 88). She also noted that she often needed a “no objection letter” from the father to do things for herself, such as to get a driver’s license (para. 93). She further expressed concern about co-parenting with the father in the UAE because she understood that as a woman she would not have equal rights (para. 109).

(b)   Second, the mother stated that she refused to return to Dubai with the children because the father would hold a grudge and the law would allow him to limit her access to her children. She feared that if she returned the father would “destro[y]” her relationship with the children, and she would be “made to feel like [she is] replaceable by -- by a maid or -- or a nanny” (A.R., vol. VI, at pp. 245-46).

(c)   Third, the mother repeatedly rejected the father’s settlement offer, which provided concrete evidence that she was not prepared to return.

(d)   Fourth, without accepting the settlement offer, the mother’s options for obtaining a residency permit are extremely limited (paras. 194-96 and 293). The mother’s precarious residency status supported her claim that she would not return.

(e)   And fifth, the mother had strong connections to Canada, including her Canadian citizenship, prior residence, supportive family ties, and property ownership in Ontario (paras. 3, 14 and 104).

[159]                     The Court does not have to speculate as to whether the trial judge ignored these considerations in rejecting the mother’s evidence. The trial judge specifically explained that he was not sure he believed the mother “for the reasons outlined earlier”. He thus did not take into account any of the considerations referred to above. His failure to address these crucial relevant considerations shows that he seriously misapprehended the evidence and made an arbitrary decision to place “very little weight” on the mother’s claim.

[160]                     The trial judge’s conclusion regarding serious harm turned on his flawed credibility determination. His error goes to the very core of the outcome of the case because it undercuts the basis for his finding about the likelihood that the children would suffer serious harm if returned to Dubai. Had the trial judge addressed the other relevant considerations in evidence before him, he would have given greater weight to the mother’s claim.

[161]                     This error also explains the inconsistent treatment of the mother’s testimony by the trial judge and the Court of Appeal. The trial judge’s ambivalence about his finding on whether the mother would return to Dubai is reflected in his later observation that it was his “fervent hope” that she would return and in his unsolicited advice: “I encourage her to consider whether it might be different [if she returns to Dubai] without the unfulfillment that came with her marriage to [the father]” (para. 380). This ambivalence continued in the decision of the majority of the Court of Appeal. The majority found no reviewable error in the trial judge placing “very little weight on the [mother’s] testimony that she would not return to Dubai” (para. 88, per Hourigan J.A.). But the majority then relied on that same rejected testimony to dismiss two of the mother’s constitutional arguments about the claimed infringements of her rights under the Canadian Charter of Rights and Freedoms , observing that “[t]he mother was quite clear in her evidence that she would not return to Dubai” (para. 175, per Brown J.A., rejecting her s. 7  Charter  argument) and stating that “the mother has made it clear that she does not intend to return to Dubai” (para. 202, per Brown J.A., rejecting her s. 2(a)  Charter  argument).

[162]                     The trial judge’s failure to consider relevant factors thus led both the trial judge and the majority of the Court of Appeal to treat the veracity of the mother’s claim of non-return inconsistently and with unacceptable ambivalence.

[163]                     My colleague interprets the trial judge’s express statements that he placed “very little weight” on the mother’s claim of non-return and “I’m not sure that I believe [the mother]” as “mak[ing] no definitive finding — whether based on her credibility or any other evidence — that the [m]other would or would not return” (para. 106). I do not read the trial judge’s reasons that way. On my reading, the trial judge’s reasons on their face rejected the mother’s claim of non-return based on his credibility finding.

[164]                     My colleague concludes otherwise by highlighting passages in which the trial judge said that the children should be returned to Dubai whether or not the mother returns, which my colleague reads as suggesting that the trial judge made no definitive finding that the mother would not return to Dubai (paras. 34, 107 and 109, citing trial reasons, at paras. 381, 387 and 389). I disagree. The comments cited by my colleague are drawn from the trial judge’s discussion of s. 40 of the CLRA, not of s. 23. In discussing s. 23 of the CLRA, the trial judge rejected the mother’s claim that she would not return based on his credibility determination (para. 368). He therefore premised his analysis of serious harm to the children under s. 23 on the basis that she would return to Dubai. His later observations under s. 40 that the children should return to Dubai, “with or without their mother” (para. 387), have no foundation in his analysis of s. 23 of the CLRA.

[165]                     In any event, if, as my colleague suggests, the trial judge did not reject the mother’s evidence that she would not return to Dubai, then this case largely turns on whether the mother has legitimate reasons for refusing to return and whether the trial judge erred in evaluating the severity of the harm to the children, issues to which I now turn.

(2)          The Mother Has Legitimate Reasons for Refusing to Return to Dubai

[166]                     I agree with my colleague that, ordinarily, a parent should not be allowed to create serious harm and then rely on it through their own refusal to return with the child if removed from Ontario (Kasirer J.’s reasons, at para. 82; Ojeikere, at para. 91). This principle against what might be called “self-engineered harm” serves to discourage child abductions (see s. 19(c) of the CLRA). I also agree that a parent’s refusal to return to the foreign country may be justified when the parent has reasonable and legitimate reasons for not returning, including a risk of imprisonment or persecution, a risk to health or physical safety (including spousal or child abuse), or a risk of a significant obstacle to employment (Kasirer J.’s reasons, at para. 83; Ojeikere, at para. 91). It follows that when a parent justifiably refuses to return, the principle against self-engineered harm does not apply. This approach avoids punishing parents who have reasonable and legitimate reasons for their claim that they will not return to the country of the children’s habitual residence.

[167]                     My colleague does not suggest that the mother self-engineered harm to the children by refusing to return to Dubai or that she must accept the father’s “with prejudice” settlement offer, failing which she must be viewed as unreasonably refusing to return (paras. 10, 13, 83, 107, 123 and 133). In my view, the mother established that she has reasonable and legitimate reasons for not returning:

(a)   The mother’s residency status in Dubai remains precarious and depends on the father’s sponsorship in securing a residency permit for her. As my colleague acknowledges, as a divorced non-national, the mother could remain in the UAE without a residency permit for only one year from the date of the divorce (para. 22). That grace period has now expired. The father obtained a divorce in the UAE in 2021, after the hearing before the Court of Appeal for Ontario, and now has both guardianship and custody (decision-making responsibility and primary parenting time) of the children in the UAE based on the UAE court’s determination that the mother had relinquished custody by wrongfully depriving the father of the enjoyment of his rights (2021 ONCA 688, 158 O.R. (3d) 565, at para. 14).

(b)   The mother was entitled to refuse the father’s “with prejudice” settlement offer that, if she returns to Dubai, he will purchase a property in her name, in trust for the children, and try to secure a residency permit for her. The mother had reasonable reasons to refuse this offer: her family and loved ones are in Ontario, including her parents and two sisters; she and Z. have regularly visited Ontario since moving to Dubai (trial reasons, at paras. 14 and 104); and she is a Canadian citizen, not a UAE national (para. 3). The settlement offer proposes that the father would purchase a house in the mother’s name, in trust for the children, to be transferred to the children when the youngest turns 18. It is understandable why the father’s offer — which would see the mother live for years in limbo as essentially a nanny under her ex-husband’s continued financial control — would not entice her back to Dubai. In any event, the father conceded, and the trial judge accepted, that he has provided no financial disclosure to substantiate whether he can buy a property in Dubai for the mother to live in (trial reasons, at para. 51).

(c)   The mother also has a legitimate basis for her concerns about living under the laws of the UAE as a woman. For instance, the trial judge accepted that in Dubai a husband is permitted to physically punish his wife without consequences in some instances (paras. 197 and 293). The trial judge also accepted that under UAE legislation a mother may lose custody of her children if she remarries (paras. 197 and 293). The mother’s evidence was that when she lived in the UAE she had to obtain a letter of permission from the father to accept an offer of employment or to obtain a driver’s license (para. 93). These circumstances contributed to the mother’s reasonable and legitimate reasons for not returning to Dubai. They are not based on generalizations and stereotypes about other countries. Rather, they are based on the findings of the trial judge and the undisputed evidence in the record about how women are treated under the law in the UAE.

[168]                     My colleague proposes to address the mother’s concerns about her precarious residency status by including as a condition of the s. 40 return order the father’s undertaking at trial to secure a residency permit for the mother (para. 133), while also acknowledging that “undertakings made by a left-behind parent before the Ontario courts may present problems of enforceability before foreign courts” (para. 13). I do not view the father’s undertaking to be an acceptable solution to the mother’s precarious residency status because the father has provided no financial disclosure about his ability to buy a second house for her to live in. More importantly, for the purposes of the prior question of s. 23 of the CLRA, the mother’s concern about her present precarious residency status is a legitimate reason for not returning to Dubai, especially now that she is divorced and has no custody or guardianship of the children under UAE law. The question at this stage is whether the mother has presented legitimate reasons for refusing to return to Dubai. In my view, she has.

[169]                     Accordingly, this is not a case in which a parent has self-engineered harm and relied on it in an unreasonable refusal to return. The mother did not rely on vague assertions that there was nothing for her in Dubai, nor did she rely only on generalized risk factors (see Ojeikere, at para. 92; Onuoha v. Onuoha, 2021 ONSC 2228, 54 R.F.L. (8th) 1, at para. 18). The mother is entitled to try to reclaim her personal autonomy and to rebuild her independent life in Canada. She provided reasonable and legitimate reasons for not returning to Dubai, reasons entitled to respect (see Gordon v. Goertz, [1996] 2 S.C.R. 27, at para. 48).

(3)          Conclusion

[170]                     I conclude that the mother established that the children would likely be harmed if returned to Dubai. I now turn to consider the severity of the harm.

B.            The Trial Judge Erred in Assessing the Severity of Harm

[171]                     The trial judge found that the children would not suffer “serious harm” if removed from Ontario (para. 370). My colleague concludes that the mother “did not demonstrate that this conclusion was unsupported by the evidence or otherwise reflects a palpable and overriding error” (para. 110). I disagree.

[172]                     The trial judge’s own factual findings regarding the expert evidence and the circumstances of these children demonstrated that the children would suffer serious harm if they were to lose their mother as their primary caregiver. The trial judge’s contrary conclusion does not attract deference because it contained material errors and failed to address the particular circumstances of these children on this record.

(1)          Children Can Suffer Serious Harm if Separated From Their Primary Caregiver

[173]                     Courts at all levels in Canada have repeatedly recognized that children can suffer serious emotional and psychological harm if they are removed from their primary caregiver, and thus it is generally in their best interest to maintain this relationship.

[174]                      This Court has often recognized the harms of removing children from their primary caregiver. For example, in Gordon, L’Heureux-Dubé J. stated, at para. 121, “[t]he assessment of the child’s best interests . . . involves a consideration of the particular role and emotional bonding the child enjoys with his or her primary caregiver. The importance of preserving the child’s relationship with his or her psychological parent has long been recognized by this Court on a number of occasions”. She added that “disrupting the relationship of the child with his or her primary caregiver will [generally] be more detrimental to the child than reduced contact with the non-custodial parent” (para. 126).

[175]                     Similarly, in Young v. Young, [1993] 4 S.C.R. 3, L’Heureux-Dubé J., dissenting in the result, noted that “the child’s relationship with the custodial parent may well be the most important factor affecting [their] long-term outcome” (p. 67, citing N. Weisman, “On Access After Parental Separation” (1992), 36 R.F.L. (3d) 35, at p. 62). As a result, “the major focus of custody decisions should be to preserve and protect the relationship between the child and his or her psychological parent” (p. 66). See also Catholic Children’s Aid Society of Metropolitan Toronto v. M. (C.), [1994] 2 S.C.R. 165, at p. 202 (highlighting a child’s bonding with their psychological parents).

[176]                      Provincial appeal courts have consistently expressed the same view. For example, in A. (M.A.) v. E. (D.E.M.), 2020 ONCA 486, 152 O.R. (3d) 81, at para. 58, Benotto J.A. accepted that Ontario courts had jurisdiction to make a parenting order involving three young children partly because they would lose the primary caregiver they had had since birth if ordered to be returned to their habitual residence in Kuwait. Similarly, in R.J.F. v. C.M.F., 2014 ABCA 165, 575 A.R. 125, at para. 70, Conrad J.A. accepted in the context of relocation within Canada that “possible harm [can] result from a change in primary parenting”. She noted that “[t]he mother has been the child’s primary caregiver for almost all of [the child’s] life, and making a change in custody would upset the parental bond that had developed between mother and child” (para. 70).

[177]                     Courts of first instance have been of the same view. For example, in Aldush v. Alani, 2022 ONSC 1536, 74 R.F.L. (8th) 113, Smith J. accepted that a primary caregiver’s legitimate refusal to return to the UAE contributed to the conclusion that the child would suffer serious harm if returned without her primary caregiver, in that case, the child’s mother (paras. 149-58). He noted that if the child were returned to the UAE without her mother, “[the child] would be deprived of the daily love, care, and support that she receives” from her mother (para. 155).

[178]                     The relationship children have with their primary caregiver is thus crucial to their emotional and psychological welfare. Separating children from their primary caregiver can affect their wellbeing and cause them serious harm. Of course, this does not mean that separating children from their primary caregiver necessarily constitutes serious harm under s. 23 of the CLRA (Kasirer J.’s reasons, at para. 78). The assessment must be fact-specific and focus on the particular circumstances of the child (paras. 72 and 75). Still, the case law’s recognition of the harm caused by separation from a primary caregiver is a useful reminder of the need for careful attention to the facts if such separation is to be countenanced.

(2)          The Trial Judge’s Conclusion on Serious Harm Is Irreconcilable With His Other Findings

[179]                     Despite the jurisprudence repeatedly recognizing that children can suffer serious emotional and psychological harm if removed from their primary caregiver, and despite the trial judge accepting expert evidence and taking judicial notice to the same effect, the trial judge found that the children would not suffer serious harm if separated from their mother. The trial judge reached this conclusion without proper regard to his own factual findings about the children’s particular circumstances. This was a serious misapprehension of the evidence and invites appellate intervention (Van de Perre, at paras. 11 and 15).

[180]                     In his brief s. 23 analysis, the trial judge found that the children would not suffer serious harm based on six conclusions:

I conclude as follows:

(i)           there is no evidence at trial that Z. and E. are in any risk of being physically harmed if they return to Dubai;

(ii)        there is some circumstantial evidence (through Ms. Parker and her opinions about infants, generally) that Z. and E. could be at risk of emotional and psychological harm if they are returned to Dubai without [the mother];

(iii)      there is no evidence at trial about the views and preferences of the children;

(iv)       there is a claim by [the mother] that she will not return to Dubai if the children are ordered to return there;

(v)         there is nothing else in the evidence at trial that this court finds to be relevant to the serious harm assessment for Z. and E. . . . ; and

(vi)       more specifically, . . . there is a total absence of any reliable evidence at trial that the court system in Dubai will do anything other than (a) determine custody in accordance with the best interests of Z. and E., if contested; and (b) award custody to [the mother], if contested, and (c) approve the settlement proposal tendered by [the father], if agreed to by the mother. [Emphasis in original; para. 366.]

[181]                     The trial judge thus accepted that there was “some circumstantial evidence” from Ms. Parker, the mother’s psychotherapist expert, that the children could be at risk of emotional and psychological harm if separated from the mother. He emphasized that, besides the factors listed, “nothing else in the evidence” was relevant to his assessment of serious harm.

[182]                     Based on Ms. Parker’s evidence, the trial judge accepted “[w]ithout hesitation” that separating an infant from their primary caregiver can cause serious harm (para. 305). He acknowledged that “infants can face serious negative effects from being removed from their primary caregiver”, but added that “I knew that before Ms. Parker testified. No trial judge needs expert evidence for that” (para. 305). The trial judge thus acknowledged that judicial notice may be taken of the uncontroversial and indisputable fact that young children can suffer serious emotional and psychological harm when they are separated from their primary caregiver (see R. v. Spence, 2005 SCC 71, [2005] 3 S.C.R. 458, at para. 53; R. v. Find, 2001 SCC 32, [2001] 1 S.C.R. 863, at para. 48). To the same effect, in ruling that Ms. Parker could give expert evidence, the trial judge stated:

. . . the general proposition that removing an infant from his/her primary caregiver may adversely impact the infant, emotionally and psychologically, would seem outside the ordinary experience and knowledge of only a trier of fact who has been living under a rock.

 

(2020 ONSC 7123, at para. 10 (CanLII))

[183]                     The trial judge accepted Ms. Parker’s expert evidence that young children can suffer short-term and long-term harm when they are separated from their primary caregiver: separation can affect an infant’s brain development and can cause cognitive impairment, aggressive behaviours, symptoms of borderline personality disorder, and difficulties managing stress. Finally, the trial judge accepted that the quality of care of a new alternate caregiver will influence how the child responds to separation from the primary caregiver (trial reasons, at paras. 307-11).

[184]                     In my view, given the case law, the expert evidence, and the facts of which the trial judge took judicial notice, the trial judge’s findings establish that the children would suffer serious harm if they are separated from their mother. Both children were infants under the age of five at the time of trial (trial reasons, at para. 3). The trial judge found that the mother has “always been th[eir] primary caregiver” and has been a “powerful force in the lives of these two children” (paras. 291 and 380). By contrast, the trial judge found that the father has “always been less involved” (para. 291). He was away from the children during working hours and travelled “quite often” in his work as a senior bank executive (para. 29). The father acknowledged that “there was a lot of time that he was not present with the children” (para. 59). The father even admitted that he has “never spent an overnight alone” with his daughter (i.e., without the mother) (para. 63). Finally, the trial judge noted that the father “has not come to Canada to visit the children since June 2020” (para. 58).

[185]                     Despite all these findings, the trial judge concluded, without explanation, that the children would not suffer serious harm under s. 23 of the CLRA if they were to be separated from their mother and returned to Dubai to live with their father. This conclusion gives rise to the reasoned belief that the trial judge must have misapprehended the evidence in a way that affected his conclusion (Van de Perre, at para. 15). Absent a misapprehension of evidence, the trial judge’s conclusion is inexplicable. The trial judge’s reasons did not apply the case law or the expert evidence to the children’s particular circumstances. In fact, in his s. 23 analysis, the trial judge did not address the children’s circumstances at all. He simply stated bald conclusions.

[186]                     My colleague suggests that the trial judge relied on the alternate care arrangement proposed by the father and the children’s return to their previous school as mitigating the harm that they would suffer if they are separated from their mother (Kasirer J.’s reasons, at paras. 104-5 and 107, citing in particular trial reasons, at para. 481). The father had proposed that if the mother did not return to Dubai, he would care for the children with help from their former nanny and two of his relatives (trial reasons, at para. 46). One relative, his older sister, lives in Pakistan with her husband and her own children (para. 52). The other relative, an aunt living in the United States, “does not know the children at all” (para. 52).

[187]                     Reading the reasons generously and as a whole, I cannot agree that the trial judge relied on such mitigation in relation to the children. The trial judge mentioned these factors at the close of his reasons in what appears to have been an attempt to persuade the mother to return to Dubai (para. 481). He did not mention these factors in making his findings under s. 23 of the CLRA (paras. 363-70). To the contrary, other than the factors the trial judge did list, he expressly found that “there is nothing else in the evidence at trial that this [c]ourt finds to be relevant to the serious harm assessment for Z. and E.”.

[188]                     In concluding otherwise, my colleague refers to the trial judge’s description of the evidence of the mother’s expert that “it would help the separation of Z. and E. from their mother if certain ‘protective’ factors were present, like returning the children home, and to where their father lives, and to a known school, and to a known nanny” (trial reasons, at para. 236, referred to in Kasirer J.’s reasons, at para. 105). But this formed no part of the trial judge’s findings regarding the expert evidence. The trial judge specifically listed his findings regarding the evidence (paras. 307-12), none of which included what my colleague now relies on. To the contrary, the trial judge stated that, other than the specific points he accepted, he “d[id] not accept any of [the] evidence [of the mother’s expert], whether given in direct examination or in cross-examination at trial, as to what will likely happen with these two children, Z. and E., if they are separated from [the mother]” (para. 312). He then reiterated that “I cannot rely upon [the mother’s expert] opinion evidence when it comes to Z. and E. specifically” (para. 315). I would therefore not rely on mitigation of the harm that the children would suffer if separated from their mother when the trial judge himself did not do so.

[189]                     The majority of the Court of Appeal also did not interpret the trial judge as having relied on mitigation of the harm that the children would suffer if separated from their mother. To the contrary, the Court of Appeal stated that the evidence of the mother’s expert — on which my colleague now relies — “provided no analysis before the trial judge on how various mitigating factors might impact the potential for harm” (para. 91).

[190]                     Even if the trial judge had sought to rely on the alternate care proposed by the father to mitigate the harm that the children would suffer if separated from their mother, it is difficult to understand how being cared for by a former nanny and two relatives who are currently outside the country — one of whom has never met the children — could adequately mitigate the harm in this case. As for the children returning to their previous school, the trial judge also did not refer to this in his s. 23 analysis, perhaps because E. was a one-year old at the time of trial and Z. was four, making any familiarity they may have had with their former nursery school environment as being of trifling significance.

[191]                     More fundamentally, I agree with the submission of the intervener Office of the Children’s Lawyer that the serious harm analysis should focus on the harm caused to the child, and not on factors that may help a child after they have been harmed (I.F., at para. 23). As L’Heureux-Dubé J. wrote in Young, at p. 85, “judges must exercise their discretion in order to prevent harm to the child” (emphasis in original). In my view, the trial judge failed to do so here.

(3)          Conclusion

[192]                     I conclude that the mother established that the children would suffer serious harm if returned to Dubai.

C.            Conclusion

[193]                     The trial judge made material errors in evaluating the likelihood and severity of the harm that the children would suffer if they are returned to Dubai. A proper application of the law to the facts — including the facts found by the trial judge himself — establishes that Z. and E. would suffer serious harm if removed from Ontario.

[194]                     Contrary to the view of the majority of the Court of Appeal, this conclusion does not create a hard-and-fast rule that serious harm will always result when an infant is removed from a primary caregiver, nor will it revive the tender years’ doctrine that children are always better off with their mother (C.A. reasons, at para. 94). The children here would suffer serious harm because of their particular circumstances, including their age, their relationships with their parents, their parents’ respective roles in their caregiving, and the father’s proposed parenting plan. The mother has always been the children’s primary caregiver and the father has always been less involved. This conclusion rests on the facts about this mother and father specifically, not on gendered stereotypes about the roles of mothers and fathers generally. Different facts may lead to a different result.

[195]                     This conclusion also respects the policy objectives underlying the CLRA of discouraging child abduction and promoting the best interests of the child. Section 23 reflects the Ontario legislature’s considered view that, in some cases, concerns about abductions must yield to the paramount objective of preventing serious harm. This is one of those cases.

III.         Disposition

[196]                     I would allow the appeal, set aside the order of the trial judge, and return the case to a different judge of the Ontario Superior Court to make a parenting order on an expedited basis, with costs to the mother throughout.

                    Appeal dismissed with costs, Karakatsanis, Brown, Martin and Jamal JJ. dissenting.

                    Solicitors for the appellant: Jamal Family Law Professional Corporation, Oakville.

                    Solicitors for the respondent: Lerners, Toronto.

                    Solicitor for the intervener the Attorney General of Ontario: Ministry of the Attorney General, Civil Law Division – Constitutional Law Branch, Toronto.

                    Solicitor for the intervener the Office of the Children’s Lawyer: Office of the Children’s Lawyer, Toronto.

                    Solicitors for the intervener the Defence for Children International‑Canada: Burrison Hudani Doris, Toronto.

                    Solicitors for the intervener the Canadian Council of Muslim Women: Lenczner Slaght, Toronto.