4.1 Introduction

‘One of the first mental frameworks which informs the intelligibility of all life’,Footnote ¹ the family seems to be one of those common concepts that are immediately understandable. Ordinary definitions of the family refer to a group of persons united by the ties of marriage, blood, or adoption, forming a single household. These persons are assumed to interact with each other in their respective positions, usually those of spouses, parents, children, or siblings.Footnote ²

Reading sociology and history, however, teaches us that the definition of the family is not as clear as it first seems. The notion of family has changed over time and its definition is the subject of intense controversy. For decades, the image of the family was based on marriage and implied a heterosexual relationship.Footnote ³ However, the narrowness of this definition, mainly based on patriarchal principles involving a male breadwinner and female homemaker,Footnote ⁴ has been increasingly criticised. Progressively the definition of family members has evolved, and in national, international, and European Union (EU) law, we find more encompassing conceptions of the family. However, the controversy is not over, namely because ‘rainbow’ and other atypical forms of family are far from being on an equal footing with the traditional nuclear family.Footnote ⁵

The purpose of this chapter is not to take a position in this debate, important as it is. Rather, its goal is to address another issue, based on the observation that the definition of ‘family’ is both uncertain and unstable in EU law. In certain parts of EU law, the family circle is limited to the nuclear family, or a slightly enlarged version of the family, while in other cases the family encompasses a broader group of persons. ‘Family’ is also defined on the basis of multiple and varying criteria. Comparing EU legal norms on free movement law, migration law, and private international law, one can only agree with Advocate General Mengozzi that the family is ‘not a uniform category’ in EU law.Footnote ⁶ Admittedly the same applies to national law, as the scope of the family varies according to the objective pursued by the legislator: to organise parental authority, to identify the legal heirs of a deceased person, to determine who are the creditors or debtors of child support, and so on. However, it is neither satisfactory nor convincing to assume that the definition of the family is purely functional in EU law.

It is not satisfactory because a totally fragmented regime does not sit well with the requirements of legal certainty which must guide the drafting and interpretation of EU law. Indeed, depending on whether or not someone is considered a ‘family member’, they may or may not have access to rights granted by EU law and be protected against breaches of other rights.Footnote ⁷ The fact that the holders of rights and obligations are designated without any apparent logic or coherence, sometimes even within the same policy area, can be considered problematic. Most importantly, the conclusion that the EU’s definition of the family is purely functional is unconvincing because it tends to assume that EU law fulfils only a very limited role in family matters, restricted by national traditions and cultures. Of course, there are reasons to believe that the EU’s influence on national models of family life is restricted. In the absence of a sufficiently homogeneous social base, the EU would lack a basis for expressing a shared European conception of the family. In addition, the EU does not ‘constitute’ families:Footnote ⁸ it only recognises the variety of affiliation, and of conjugality that we find in the Member States’ legislation,Footnote ⁹ hence building upon the families constructed at national level.Footnote ¹⁰ However, this reading is contradicted by the different works that, despite reaching opposite conclusions, all agree that EU law influences family patterns in Europe. Whether they argue that EU law ‘undermines the family as a legal institution’,Footnote ¹¹ that the EU creates ‘forms of family life’,Footnote ¹² or that the contribution of the Court of Justice of the EU (CJEU) to the protection of rainbow families is too weak,Footnote ¹³ EU law scholars are increasingly united around the idea that the EU’s influence on national models of the family is real.Footnote ¹⁴

This chapter takes this idea as a starting point. It aims to show that, under the heterogeneous approach to the family, there is a rationale underpinning the EU’s – apparently – variable and unstable approach to the family. It seeks to describe how the notion of ‘family member’ contained in EU legal provisions reflects the EU’s perceptionFootnote ¹⁵ of what European society is (and what it should be). The notion of ‘doing family’, borrowed from sociology, will help to account for the reasoning of the EU legislator and judge. In EU law, there is no unitary definition of what it is to be a family member, but EU legal norms often tell us what it means to act like a family member. Some people, who cannot claim to have conjugal or parental bonds with rights’ holders, are nevertheless referred to as ‘family members’ on the sole ground that they are behaving like family members. In characterising such people as family members, the EU values the family functions they are performing: protecting, caring, educating, and supporting the social integration of their family. However, this chapter will show that ‘doing family’ is a social fact, an objective social reality: it does not equate to ‘feeling like’ a member of the family. While the EU does not ignore feelings and the willingness to form a family, it relegates them to second place. This preference for a definition of family that is based on the social functions performed rather than on any self-definition is a political and ethical choice. The Union does not so much value the feeling of belonging to the family as the role that the constitution and preservation of family life plays in the service of European construction.

This chapter is divided into six sections. Sections 4.2 and 4.3 substantiate the idea that family is a variable geometry figure in EU law. Section 4.2 describes the variable scope of the family circle in EU legal instruments, and Section 4.3 shows that ‘family’ is a missing category of EU law. Thereafter, the chapter seeks to identify the rationale underpinning the delineation of the family members by EU legal norms. Section 4.4 presents the notion of ‘doing family’, borrowed from the work of David H. J. Morgan,Footnote ¹⁶ who focused on family practices. It shows the value of this notion for understanding the rationale of the EU approach. Section 4.5 invites nuance in the use of the notion of ‘doing family’ to describe EU law. The EU values the fact of ‘doing family’ but, in contrast with Morgan’s work, there is no room for the self-definition of family members. Forming a family supposes that its members perform social functions, like providing care or contributing to the social integration of its members. The rationale for this limitative definition of the family is explained in Section 4.6: family is a form of social membership that enables people to be part of a broader community: the European society. EU law models the definition of EU families on the EU project to construct a cohesive society. Section 4.7 concludes that despite the fragmentation of sources and the modulation of family circles, EU law projects a certain representation of the family. Even when unnamed, ‘European families’ exist.

4.2 The Variable Scope of the Family Circle

To illustrate the impossibility of accurately describing the ‘European family’, Strumia resorts to the notion of the ‘variable geometry family’. The EU institutions rely ‘on a flexible, pragmatic idea of family that leaves potential room to several models of cohabitation and reciprocal responsibility, and to a variety of underlying bonds, from the biological, to the legal, to the factual and affective’.Footnote ¹⁷ Apart from the members of the nuclear family – the spouses and children – who are systematically included in the group of ‘family members’, the precise boundary of the family circle is unstable and difficult to assess in EU law.

Several factors generate this instability. Time is the primary factor of variation. Since the 1950s, the definition of the family circle has evolved, espousing some of the social and legal changes in the Member States. This is visible in free movement law. Regulation 1612/68Footnote ¹⁸ on workers’ mobility restrictively defined the family as being composed of the sole ‘spouse‘, children, and the dependent relatives in the ascending line of the worker and hisFootnote ¹⁹ ‘spouse’. Three decades later, the 2004/38 DirectiveFootnote ²⁰ included in the family circle the ‘partner’ with whom the Union citizen (a man or a woman) has contracted a registered partnership. Following the same logic, siblings and relatives who were not mentioned in the 1990 Dublin ConventionFootnote ²¹ have been counted as members of the (extended) family by ‘Dublin III’ Regulation 604/2013.Footnote ²² The proposal for an Asylum and Migration Management Regulation confirms that the EU institutions are willing to retain this broad definition of the family circle.Footnote ²³

However, change over the course of time is not the only source of complexity. Even when adopting a static rather than a dynamic reading of the norms dealing with family life, it is hard to form a clear picture of the family. The fragmentation of legal sources is a second factor preventing the emergence of a uniform category of family in EU law. The perimeter of the family varies significantly from one text to another, even within one and the same policy field. A typical example is that of unmarried partners. Directive 2013/33 on the reception of applicants for international protectionFootnote ²⁴ is inclusive: its Article 2(c) designates as a family member ‘the spouse of the applicant or his or her unmarried partner in a stable relationship’. Accordingly, Article 2(g) of the Dublin Regulation mentions the unmarried partners engaged in a durable relationship. Yet other texts, like the Family Reunification DirectiveFootnote ²⁵ and the EU Citizens’ Directive, are far less inclusive. They rely on a distinction between two groups of people, one named ‘family members’ which corresponds to the (slightly extended) traditional nuclear family, and the other one composed of people having other types of family ties. Unmarried partners fall into this second group.Footnote ²⁶ The same logic applies for siblings and relatives. In most EU legal instruments, they are not included in the family circle. Article 6 of the Dublin III Regulation is an exception: it refers to a heterogeneous group composed of ‘family members, siblings or relatives’ of the unaccompanied minors. ‘Relatives’ are defined by Article 2(h) as ‘the applicant’s adult aunt or uncle or grandparent who is present in the territory of a Member State’.

These lists of family members may appear to have been composed at random and without any logic because the links between family members are very diverse. Parental and conjugal ties are predominant, but often EU law also recognises as family members people who interact from other circles and relationships. A recent example is that of the Council Implementing Decision 2022/382 of 4 March 2022 establishing the existence of a mass influx of displaced persons from Ukraine.Footnote ²⁷ In addition to the spouse, the unmarried partner in a stable relationship, and minor unmarried children, Article 2 of the Decision designates as persons who shall be considered to be part of a family ‘other close relatives who lived together as part of the family unit at the time of the circumstances surrounding the mass influx of displaced persons, and who were wholly or mainly dependent on the sponsor’. The ‘other close relatives’ is a vague category that serves to include, within the scope of the family, persons not having any predetermined legal family links but who depend on the sponsor. It will be remembered (from the previous paragraph) that the same category is mentioned in the Dublin Regulation but there ‘relatives’ refers to a different category of people: ‘the applicant’s adult aunt or uncle or grandparent who is present in the territory of a Member State’ (Article 2(h)).

These examples, which attest to the categorical diversity of family in EU law, point to a third factor of variation in the definition of the family: the objective of the law. The family circle is defined according to the purpose of the EU norm under consideration. When the legislator’s objective is to confer a legal status and to grant rights to individuals, the family circle tends to be narrowly defined. This is particularly apparent in the norms that distinguish between two categories of family members, who are accorded distinctive rights and treatment. Following its Article 3(1), the EU Citizen’s Directive, which confers several important rights, ‘shall apply’ to the ‘family’ which is composed (under Article 2) of the spouse and registered partner, children, and ascendants in direct line. In contrast, under Article 3(2), the host Member States are only required to ‘facilitate’ entry and residence for the ‘other family members’ who are the persons not falling under the definition of Article 2 who are dependants or members of the household of the Union citizen. Accordingly, and despite the fact that he/she is not even named a ‘family member’, the ‘partner with whom the Union citizen has a durable relationship, duly attested’ is mentioned in Article 3(2)(b) as a person who ‘may enjoy’ facilitated access to entry and residence. The legislator has opted for the same approach to family in concentric circles in the Family Reunification Directive. It distinguishes between a ‘first family circle’ including the sponsor’s spouse and the minor children of the sponsor and spouse, composed of persons who have a right of entry and residence in the host State; and a ‘second family circle’ comprising first-degree dependent relatives in the direct ascending line of the sponsor or spouse, and adult unmarried children of the sponsor or spouse, composed of persons whose entry and residence is dependent on the host State being willing to welcome them.

These examples also show that the delineation of the family varies depending on the degree of harmonisation the legislator seeks to achieve. When adopting the EU Citizen’s Directive, the legislator clearly sought to limit the Member States’ discretion; hence the compilation of a rather precise and closed list of family members. In contrast, in the field of social security, preference was for a margin of appreciation left to national authorities. Logically, Article 1 of Regulation 883/2004Footnote ²⁸ defines the ‘member of the family’ as ‘(i) any person defined or recognised as a member of the family or designated as a member of the household by the legislation under which benefits are provided …’.

It is clear from these examples that the definition of family – and the protection of family life – is often instrumental to achieving other objectives. This is not a novelty. Caracciolo di Torella and Masselot have argued that ‘the regulation of the family has not been an end per se, but has been subsidiary to the successful completion of the “main building”, namely the European Community (EC) common market. As a result, both soft law and binding measures were a bundle of ad hoc measures which did not support a coherent approach within this area.’Footnote ²⁹ Given the multiple objectives pursued by the EU institutions in the different areas of EU law, the prospect of a clear and a priori definition of the family common to the different branches of EU law appears somewhat illusory.

4.3 The Family: A Missing Category of EU Law?

It is not unreasonable to affirm that ‘family’ is an elusive category of EU law. Terminology confirms this idea. In the different secondary law instruments that organise family life, in particular in social law, free movement law, or private international law, the entity ‘family’ is rarely mentioned as such. The preferred terms of the norms are ‘family links’, ‘family relations’, or ‘family members’, leading some scholars to consider that European law is constructed on the basis of, and for, individuals alone.Footnote ³⁰ Some go as far as concluding that the ‘European family’ does not exist, at least in its traditional sense of a social institution.Footnote ³¹

This description fits well with the analysis of free movement law, a domain in which the logic is ‘of dissuasion or deterrence’.Footnote ³² Free movement law, Advocate General Bobek argues, is built on the premise that the Union citizen will be discouraged from moving, ‘as those personally close to him will be barred from joining him’.Footnote ³³ Given that social perceptions are changing and that there is an increasing range of forms of cohabitation, the definition of the persons who are ‘close’ to the Union citizen is not – and cannot be – defined a priori. Sometimes, the deterrence effect is even stronger with regard to a partner under Article 3(2) of Directive 2004/38 than it is with regard to some family members listed in Article 2. This is why Advocate General Bobek adds: ‘I am simply suggesting that with regard to who is effectively “close” to a person, formal box-based generalisations are hardly appropriate.’Footnote ³⁴

It may be the same reluctance vis-à-vis ‘formal box-based generalisations’ that guided the legislator when adopting Regulation 2201/2003 concerning jurisdiction and the recognition and enforcement of judgments in matrimonial matters and matters of parental responsibility.Footnote ³⁵ The text does not specify who are the persons holding parental responsibility and having the rights of access to the child defined in Article 2(10). Whether certain persons, such as grandparents, can be considered family members and have rights of access to their grandchild is not explained. This lack of clarity results from the negotiations of the time.Footnote ³⁶ The legislator, pondering who may exercise parental responsibility, considered several options: limiting the persons concerned to one of the parents of the child or, conversely, imposing no limitations on specific persons whatsoever. Ultimately, the legislator opted not to provide for a limitation of the range of persons who may exercise parental responsibility. For Advocate General Szpunar, we can even infer from the legislator’s silence that its goal was to extend the scope of Regulation 1347/2000.Footnote ³⁷ The legislator contemplated all decisions concerning parental responsibility and rights of access ‘irrespective of the nature of the persons who may exercise those rights’Footnote ³⁸ and without excluding grandparents. There is no predetermination of what family could mean because flexibility was the preferred option. The legislator thus achieved the paradoxical outcome that, in a text governing parental responsibility, there is no trace of a definition, or even a conception, of the family.

The semantics employed by the different texts of EU law bring to light the conceptual confusion surrounding the notion of family. Different terms (‘family members’, ‘other family members’, ‘relatives’, ‘other relatives’, ‘siblings’, ‘siblings and parents’, ‘parents’) coexist to describe the persons having ties with the EU citizens or sponsors or person to be taken care of, often in an inconsistent manner. True, the rationale is to modulate the rights granted to the different categories of persons. But this mode of presentation and naming creates a conceptual ambiguity: Is it coherent to talk about a ‘family’ when some people are not even titled ‘family members’Footnote ³⁹ and others, named ‘other family members’, remain outside the scope of the provision that defines ‘the family’?

Practical problems also arise from the inconsistent reference to the notion of ‘family members’. The Banger caseFootnote ⁴⁰ illustrates the difficulties arising from the inconsistency in Directive 2004/38. Article 3(2) of the Directive grants a facilitation regime of mobility for the unmarried partner with whom the EU citizen has a durable and duly attested relationship but does not set out the specific procedural guarantees that shall accompany its implementation. Therefore, the judicial guarantees to be applied to unmarried partners, who are not on the list of ‘family members’ in Article 2, remain unclear. A difficulty emerged because Article 15 of the Directive on procedural guarantees refers to Union citizens and their ‘family members’ with no other indication. Both the Commission and the UK government argued that the notion of Union citizens and their ‘family members’, which is legally defined in Article 2(2) of the Directive, does not include extended family members. Advocate General Bobek and the Court took another position. The Advocate General noted that the notion of ‘family members’ does not appear to be used consistently in the provisions of Directive 2004/38.Footnote ⁴¹ Since under Article 15 procedural safeguards shall apply ‘to all decisions restricting free movement of Union citizens and their family members’ without referring to Article 2, it is not unreasonable to consider that the procedural guarantees to apply to family mobility should not vary according to the type of conjugal bond. Following the same logic, the Court held that the procedural safeguards provided in Article 31(1) of Directive 2004/38 are applicable to the persons envisaged in Article 3(2)(1) of the directive. This approach must be welcomed insofar as procedural guarantees are accorded to members of the extended family. Yet it is not unreasonable to question the Court’s logic, which assumes that a unitary category of family exists, despite the distinction that Directive 2004/38 makes among family members. At this stage, the legitimate question to be asked is whether the Court and the other EU institutions have a vision of the ‘EU family’ in mind. Is there a rationale underpinning their approach? These questions can be answered in the affirmative, provided that due consideration is given to the notion of ‘doing family’.

4.4 Defining Family Members as Those Who ‘Do Family’

Until now, the family has been described in this chapter as a category without substance. It is now time to take a new lens: instead of who is in the family or what is a family, we might instead ask what makes a family.Footnote ⁴² For Morgan,Footnote ⁴³ family is a facet of social life: it represents a quality rather than a thing. Morgan’s work is useful for our analysis because he has shifted sociological analyses from family as a structure to which individuals belong, toward understanding families as sets of activities, which take on a particular meaning, associated with family. This line of reasoning is insightful because it has proven to be quite impossible to find a unitary family structure in EU law. In EU legal norms, in contrast, one can identify different traces of ‘family activities’ and ‘family practices’. It appears from various legal provisions that the EU recognises situations in which individuals ‘behave’ and ‘act as family’ members. Often the individuals are granted rights when, or because, they perform one of the main social functions which are generally assigned to the family: to provide care, protection, and education, and to contribute to the social integration of its members.

True, the social functions generally assigned to the family are not valued equally. ‘Doing family’, in EU law, relates primarily to performing a caring function. The Court formulated this idea implicitly in the Valcheva case. Having noted that the legislator had left it undecided who could be a ‘holder of parental responsibility’, the Court then provided a defining criterion of ‘parents’: they are the persons ‘with whom it is important for the child to maintain a personal relationship’.Footnote ⁴⁴ In so judging, the Court has not emphasised a specific type of bond which would define a parent; a case-by-case analysis will be required to identify who the ‘parents’ are. Nevertheless, the Court has chosen to emphasise the importance of bonds in the interest of the minor.

In most free movement and migration norms, the caring function plays a more explicit role. Article 3(2) of the EU Citizen’s Directive defines as ‘other family members’ the persons who ‘are dependent or members of the household of the Union citizen having the primary right of residence, or where serious health grounds strictly require the personal care of the family member by the Union citizen’. The same logic guides the Dublin III Regulation, the Asylum Directives,Footnote ⁴⁵ and the 2015 Relocation Decision:Footnote ⁴⁶ people with no biological bonds and who are not de jure family members are nevertheless referred to as family members when they provide, or need, care. The most explicit provision is Article 4(2) of the Family Reunification Directive, which allows the Member States to authorise the entry and residence of: the first-degree relatives in the ascending line of the sponsor or his/her spouse, where they are dependent on them and do not enjoy proper family support in the country of origin; the adult unmarried children of the sponsor or spouse, where they are objectively unable to provide for their own needs on account of their state of health; the unmarried minor children as well as the adult unmarried children of the partner, when these children are objectively unable to provide for their own needs on account of their state of health. The same logic guided the Council when adopting Implementing Decision 2022/382 which triggers the application of the Temporary Protection Directive for people from Ukraine.Footnote ⁴⁷ In this text, the family includes the ‘other close relatives’ who lived together as part of the family unit ‘and who were wholly or mainly dependent on the sponsor’.

The notions of ‘need’ and ‘care’ take on such importance in EU law that they have even led the Court, in the Depesme and Kerrou case,Footnote ⁴⁸ to redefine the notion of ‘children’ and ‘parent’, despite the strict meaning of the notions under Regulation 492/2011.Footnote ⁴⁹ Ms Depesme, Mr Kauffmann, and Mr Lefort each lived in a reconstituted family unit – living in France and Belgium – consisting of a biological mother and a stepfather. Each of them applied for a study grant on the basis that their stepfathers had worked in Luxembourg continuously for more than five years. Yet the Luxembourgish authorities refused their applications on the grounds that they were not legally the ‘children’ of frontier workers but merely the ‘stepchildren’. The Luxembourgish government argued that the concept of ‘child’ had to be interpreted restrictively.Footnote ⁵⁰ Following the AG, the Court confirmed the social value of providing maintenance and care. It held that the parent–child relationship has to be defined not in legal but in ‘economic terms’, in that the child of a step-parent with the status of migrant worker can claim a social advantage where the step-parent contributes to the child’s maintenance. The centre of gravity of the judgment is not about membership or legal linkages, it is about performing a family role, day after day. ‘Doing family’ is a notion that comes in via the court’s application to complement de jure family bonds.

On closer inspection, it appears that this conception has been at work for many years. Already in the Carpenter case, the Court had the concept of care play a decisive role.Footnote ⁵¹ Accordingly, in Ibrahim and Teixeira,Footnote ⁵² the Court conferred a right of residence on a mother of a child in education on the basis of her being the primary carer for the child. In Teixeira, the Court even considered that the right of residence in the host Member State of the parent who is the primary carer for a child of a migrant worker ends, in principle, when the child reaches the age of majority unless the child continues to need the presence and care of that parent to pursue and complete his/her education.Footnote ⁵³ Later, in Alarape,Footnote ⁵⁴ the Court confirmed the possibility of the primary carer remaining on national soil as long as the child needs care. In so judging, the Court went beyond the limits of EU secondary law, which did not ground a right of residence on the sole fact of being a mother.Footnote ⁵⁵ ‘Acting like a parent’ is what the Court takes into consideration. It serves either to replace – as in Depesme and Kerrou – or to complement – as in Alarape – de jure family ties.

In sum, when looking at EU secondary law and at CJEU case law through the lens of ‘doing family’, a more coherent picture of the family emerges. The ‘family members’ remain identified by the conjugal or kinship bonds they share with EU citizens or migrants. The extended family members, whether named ‘other family members’, ‘other relatives’, ‘relatives’ or unnamed altogether, are mainly identified by the family functions they fulfil or benefit from actively caring and protecting, or passively needing care and protection. This approach, however, is an objective social reality and ‘doing family’ should not be confused with ‘feeling like a family’.

4.5 ‘Doing Family’ as an Objective Social Reality

When the Coman judgmentFootnote ⁵⁶ was handed down by the Court, very contrasting comments were made by EU legal scholars. Since the Court held that the mobility rights conferred by the EU citizenship Directive shall be granted to the same-sex spouse of an EU citizen, even without the requirement that the host Member State has established same-sex marriage, the judgment has been described as bold and progressive.Footnote ⁵⁷ It was indeed legitimate to infer from the judgment that the Court was willing to interpret broadly the concept of EU citizens’ family members,Footnote ⁵⁸ and that the EU was progressively accepting new forms of families. However, a far less enthusiastic reading of the case emerged from sceptical scholars who denied that the Court had transformed the definition of family members. The outcome, they argue, was ‘simply mandated by the language of the relevant legal provisions’ since their inception, given that the term ‘spouse’ was ‘purposefully gender-neutral in the Directive 2004/38’.Footnote ⁵⁹ When families composed of a same-sex couple move to a Member State that does not recognise same-sex couples, they remain confronted with the possibility that the Member State will refuse to legally recognise the familial ties among the family members as these have been legally established elsewhere. The Coman ruling indeed creates an obligation to recognise such marriages but ‘for the sole purpose of granting a derived right of residence to a third-country national’.Footnote ⁶⁰

This critical reading rightly underlines that even in Coman, a judgment that enshrines a more liberal view of the family, the family remains defined by stringent objective criteria. The Court in no way promotes a definition of the family that is self-defined or ‘rooted in individual biographies’.Footnote ⁶¹ This contrasts with Morgan’s approach which inferred from the notion of ‘doing family’ that social actors creatively constitute their own social world and that ‘an individual’s understanding of “my family” is subject to change over time’.Footnote ⁶² Morgan emphasises the fluidity, the diversity, and the multifacetedness of the family. Here lies the limit of Morgan’s theory for the analysis of EU law. In EU secondary law and in CJEU case law, ‘doing family’ and ‘feeling part of a family’ are two separate things. ‘Doing family’ remains limited to performing a set of activities or actions: it is a matter of how people behave, not how they feel, how they think of themselves as a social group, or how they love.

In CJEU case law, the motivations or the reasons for ‘doing family’ are irrelevant. This appears clearly from the Reyes case,Footnote ⁶³ in which the Court decided that for a twenty-one-year-old daughter of a Union citizen to be regarded as being a ‘dependant’ of that citizen, the existence of a situation of real dependence must be established. That status can only be the result of a ‘factual situation’ characterised by the fact that material support for that family member is provided by the Union citizen who has exercised his right of free movement or by his spouse. The Court underlines that there is no need to determine ‘the reasons’ for that dependence or for the recourse to that support. The same rationale applies in Depesme and Kerrou. The national judge sought to clarify the extent to which emphasis had to be placed on the fact that the frontier worker ‘continues to provide for the student’s maintenance’ without necessarily being connected to the student through a legal child–parent relationship, where a sufficient link of communal life can be identified. The Court replied that the status of dependent member of a family is ‘the result of a factual situation, which it is for the Member State and, if appropriate, the national courts to assess’. The status of family member can be evidenced by ‘objective factors’, such as a joint household shared by that worker and the student. It is not necessary to determine the ‘reasons for the frontier worker’s contribution’ to the maintenance of the student or make a precise estimation of its amount. ‘Doing family’ is an objective fact and the willingness to act as a parent is secondary. Such willingness can be inferred from the acts of ‘parents’ but needs to be proved.

Of course, the Court is not unaware of the emotional side to family life but feelings and emotions are not sufficient to ‘constitute’ the family. This is apparent in AlarapeFootnote ⁶⁴ – a case in which the national judge was asking whether, for a parent to qualify as a ‘primary carer’ so as to derive a right of residence from a child under Regulation 1612/68, it is necessary for that child to be dependent on such a parent, to reside in that parent’s household, and to receive emotional support from that parent. The Court held that determining whether an adult child does or does not continue to need the presence and care of his parent supposes one should:

Emotional ties are not ignored but they mainly serve to establish the children’s dependency and are not central to the definition of family ties.Footnote ⁶⁶ In other words, it seems that recognising the importance of ‘doing family’ does not equate to giving individuals the choice to decide for themselves who counts as family members. Rather, because the family is the ‘basic cell’ of European society, its definition is modelled after the EU’s goal to create a cohesive society.

4.6 The Role of ‘Doing Family’ in the Design of European Society

Among EU law scholars, the story increasingly told is one of the EU making space for greater individual emancipation and self-determination but leading to a society in which the sense of community and membership is frustrated.Footnote ⁶⁷ This reading is well established among those who believe in the negative influences of the EU on family law. The EU supposedly destabilises the family order, accentuating the tendency towards individualism and the break-up of the family institution. The Coman judgment, and the recent Pancharevo ruling,Footnote ⁶⁸ may have increased the impression that in promoting subjective rights and overvaluing the interest of the individual at the expense of the socially dominant conception of the family, the Court contributes to the overall phenomenon of relegation of the family as a legal institution.

There is, however, another narrative to be taken seriously. It tells that EU law strives to convert European individuals into members of the social and institutional spheres that prevail in their Member State of residence. The family is one of the ‘collective entities’ which individuals are destined to belong to.Footnote ⁶⁹ Azoulai argues that family membership is of particular importance because it is ‘a proxy for social membership’:Footnote ⁷⁰ families are not ‘mini-societies’ but entities which are central to the construction of a cohesive European society, hence the importance accorded by the EU to situations in which individuals ‘do family’. Caring and protecting, participating in the construction of stable communities are all some of the individual acts that contribute to the construction of a broader social community.

Different factors support this social membership narrative. It can be observed that in various provisions and judgments, family members are encouraged to play two main roles that are decisive for the cohesion of European society. The first role is to provide material and emotional support to the most fragile and vulnerable members of the society. This aspect is central in the rules on the family reunification of refugees. Special attention is given to the possibility for refugees to recreate a family: being vulnerable people, they need to enjoy a sphere of solidarity. This is why, in the TB case,Footnote ⁷¹ and in the A.S. case,Footnote ⁷² the Court has restated the importance of family reunification as a necessary way of making family life possible, ‘which is of particular importance for refugees on account of the reasons which obliged them to flee their country’.Footnote ⁷³ Children being fragile and vulnerable individuals, the Court also values the role of their caregivers. As mentioned earlier, this led the Court in Depesme, Teixeira or Chavez Vilchez to characterise the stepfather as a ‘parent’ despite the absence of legal links, solely on the basis of the assistance provided. In the SM case,Footnote ⁷⁴ the Court also decided that a child under a legal guardianship system such as Algerian kafala, although cannot be regarded as a ‘direct descendant’ of a Union citizen for the purposes of Directive 2004/38, can be defined as one of the ‘other family members’ referred to in Article 3(2)(a).Footnote ⁷⁵ The words employed in Article 3(2)(a) are indeed capable of covering the situation of a child placed with citizens of the Union under the kafala system ‘in respect of whom those citizens assume responsibility for its care, education and protection’. In these different cases, the family was always conceived as the social entity best suited to provide protection and adequate care for the most fragile individuals.

Family members are called on to perform a second social function: to contribute to the stability of social relations. Coman is an interesting example because the argument which grounded the judge’s decision to grant same-sex spouses the right to move in the EU is the importance of protecting the family life that was created ‘or strengthened’ in another Member State. In so doing, the Court emphasises the individual’s need for stability, a social reality that is expected to benefit both individuals and society. In other cases, the Court values the actions taken by ‘family members’ which ensure the continuity of children’s education and integration in the host society. This is noteworthy in cases in which mothers (as in Chavez Vilchez) or stepfathers (as in Depesme) stand in for the children’s defaulting father by paying for maintenance and education in his stead. These persons deserve to be called ‘family members’ because giving children the opportunity to attend school and pursue further education is ‘a condition for the best possible integration of children of migrant workers in the social life of the host member State,’ Advocate General Kokott explainedFootnote ⁷⁶ The mother–child or stepfather–stepchild tie is not reducible to a form of universal human relation, based on legal ties or on feelings. It is based on specific relational links that contribute to the European project of creating a cohesive society. ‘Family members’ can thus be described as the active persons whose presence and care are required for the purposes of integration and stability. They are defined by the specific role they are destined to play in the common interest. Being a family member is a role given to the individual as part of broader relations with others embedded in European society. Performing useful social functions is what makes a family in EU law.

4.7 Conclusion

It is very tempting to believe that EU law fulfils only a limited role in family matters, restricted by national traditions and cultures. There are good reasons to believe that the (possible) influence of the EU on national models of family life is limited. In the absence of a general EU competence clause, there can be no real family law. In the absence of a sufficiently homogeneous social base, the EU would lack a basis for expressing a shared European conception of the family. In other words, the EU would not substitute its own value and representation.

However, this chapter has revealed that this description fails to account for the power of the EU’s language which contributes to modelling ‘the way we live and conceive our lives’.Footnote ⁷⁷ It has been shown that despite the apparent variable geometry of the family, and the absence of a uniform category of ‘family’, the way in which the EU characterises a person as a ‘family member’ obeys a form of logic and expresses a certain rationality. Despite the fragmentation of sources and the modulation of family circles, rights and obligations are not randomly distributed; EU law projects a certain representation of the family. We have observed that in addition to the de jure family members, other persons are counted as family members on the basis of their ‘behaving’ like family members. Even when unnamed, European families exist: they are the groups of people who are assumed to perform – or asked to prove that they do perform – different functions, like education, care, protection, and socialisation. These roles are central because they contribute to a broader ambition – participating in the cohesion of the whole of European society, which is composed not of isolated individuals, but of social groups related by strong human bonds, and not only economic ties. This can be viewed as a limited contribution to the emergence of European society or, on the contrary, as the sign that EU integration is progressively taking on a more human face.