Ties that bind and unbind: charting the boundaries of European Union citizenship

ABSTRACT Acquiring or losing European Union (EU) citizenship is contingent on the possession of citizenship of one of the Member States. Since the early 1990s, scholars have debated the unique character of EU citizenship. Remarkably, these debates have paid scarcely any attention to how EU citizenship is (un)recognised through the different conceptions of membership in the 27 countries that make up its boundaries. This article argues that understanding the substance of EU citizenship requires a look into the different domains of citizenship laws in each of the Member States. We present a novel conceptual framework for studying citizenship regimes through four types of citizen-state links: lineage, territory, sponsorship and merit. We find that the disparity among the Member States in who is (un)seen as an EU citizen results from the different ways in which the four types of state-citizen links are articulated in the rules for citizenship acquisition and loss.

through the diverse connections that exist between individuals and the Member States. These 'citizenship links'as we refer to the conceptions of legal grounds for (un)recognising individuals as nationalsvary across countries and change over time. They are ingrained in the mechanisms of citizenship acquisition and mirrored in the modalities through which citizenship can be lost. The uneven edges of EU citizenship have thus been created by the differentiated governance of citizenship in a multi-level structure, where national citizenship is the exclusive gateway to the supranational status.
To show these dynamics, we root our analysis in citizenship studies. Rather than studying EU citizenship per se, we look at how the coming together of different citizenship links that exist across countries shape its boundaries. We commence by constructing a typology of the domains of citizenship law to demonstrate how accessing or losing citizenship is rooted in a particular logic of 'link'. We then deploy the typology to study the regulation of citizenship in the 27 EU Member States as of 1 January 2022. We deliberately focus on the substantive dimension 'as it is', as the temporal one is uneven by default. In our study, we depart from the GLOBALCIT Citizenship Law Dataset (GLOBALCIT 2021) to identify modes of acquisition and loss constitutive of each of the four types of recognised connections between states and individuals. We then analyse the respective legal provisions relying on the Global Nationality Laws Database (GLOBALCIT 2022a(GLOBALCIT , 2022b. Finally, we supplement our understanding of the legislative framework with insights from national experts on citizenship law in each of the 27 EU Member States, who havebetween December 2020 and June 2021provided us with detailed information on legislative changes in their respective country. All information has been obtained by means of personal correspondence, whereby experts have been asked to fill out a questionnaire referring to specific modalities for the regulation of citizenship (e.g. years to naturalisation by marriage; other conditions for naturalisation by marriage).
Our work stands to make a contribution to both citizenship and European studies. It adds to the work of those who study citizenship by providing a new lens for studying the kinds of putative links between individuals and states. Unlike previous typologies, which have focused primarily on the grant of citizenship, our framework proposes that the same grounds for who is (un)seen as a citizen are contained in the rules governing both acquisition and loss. Such a lens allows us to highlight the significance of variation within states for setting the parameters of citizenship beyond the state. We thus apply this model to empirically study the 'formation' of supranational EU citizenship by unpacking the differentiated governance of membership in this multilevel structure. We build on the recent wave of scholarship questioning whether the allocation and loss of nationality among the Member States should be subject to some minimum common standards for a 'genuine link', and if so, what would constitute such a 'link' and what would make it 'genuine' (European Commission 2019; van den Brink 2022; Van den Brink 2018;Wagner 2021;Maas 2021). While the purpose of this article is not to disentangle the normative conundrum of what standards should be set for citizenship acquisition and loss to solve problems of over and under inclusion, our analysis highlights where there are already existing norms among the Member States, identifying places of agreement on procedure or substance as to what constitutes a widely acceptable connection. Equally, our study identifies the outlier 'links' that are contested as membership boundaries. As such, it opens an intellectual space for an empirically grounded debate on the conditions for recognising and challenging the rules for citizenship acquisition and loss.
Framing citizenship law domains: a typology of recognised state-citizen links Our conceptual framework starts from the premise that the boundary of a state's membership is in part shaped by the legislation that governs the status of 'citizen'. The literature on citizenship and state boundaries has previously focused on whom the state can exercise the legitimate use of power (Bauböck 2017;Abizadeh 2008;Miller 2009). Our approach unpacks the different legal bonds that draw the boundary of the state and defines who becomes included in 'the people' of the Member State and consequently, the EU.

The four types of links engrained in citizenship laws
We suggest that the spheres of legal membership are built up of four types of links that are constructed through the different grounds for citizenship acquisition and mirrored through those for citizenship loss. Starting from the GLOBALCIT modes typology, which outlines 27 ways citizenship can be acquired and 15 ways citizenship can be lost, we reconceptualise these grounds by their underlying logic of membership. This was done using an inductive approach, by uncovering and grouping the modes by the main condition required for acquisition or loss of citizenship. Thus, we found that this occurs through four categories: (1) lineage; (2) territory; (3) sponsorship; and, (4) merit. Such a categorisation of links allows us to capture variation within states as well as between them by unravelling the 'constellations' that tie together the citizen, the state and the transnational space (Bauböck 2010). We begin by conceptualising the content, historical roots and evolution of each type of link. Our typology unpacks the logics of membership policies, how they relate to one another, and how they shape the boundaries of a polity.
First, lineage-based links are conceived either as (1) biological ties operating as birthright (ius sanguinis) or (2) connections to the state through presumed lineage derived from 'perceived common origins or descent', shared culture, language, or religion (Joppke 2005, 3). While blood-based membership transmission can be traced back to the ancient Greek city-states (Riesenberg 1992), the contemporary renditions of ius sanguinis are rooted in the French Civil Code of 1803, which then spread through Europe (Weil 2011). Traditionally, ius sanguinis was a feature of civil law countries; however, most states have such a provision in their current citizenship law. Birthright citizenship through ius sanguinis is devised to ensure the intergenerational continuity of a population linked to a state (Vink and Bauböck 2013;Bauböck 2018). Citizenship transmission through parentage represents an automatic establishment of a 'link', which can be materialised within or outside the country's territorial borders. Beyond descent, putative lineage also constitutes a widely recognised connection. This is operationalised in citizenship law by facilitating naturalisation for those with a presumed shared cultural heritage, imagined future, or as a remedy for past wrongdoing based on ancestral history (Maatsch 2011). Lineage-based links have the purpose of preserving or reconstituting a community through consanguinity. Presumed lineage may be rooted in both ancestry or supposed affinity irrespective of the place of residence (Harpaz 2019;Ragazzi and Balalovska 2011;Dumbrava 2019). Christian Joppke has pointed out that immigration policies have had an 'inclination in modern nation-states to select newcomers in light of their proximity to the particular ancestry' and the same can be said for many states' citizenship laws (Joppke 2005, ix). In some states, a 'link' based on lineage can be lost after a certain number of generations born abroad, while in others the transfer is unlimited.
Second, recognised links can reflect the relationship between an individual and the physical territory of a state. They encompass the ideas that (1) territory itself and (2) residency within a bounded place creates a bond between a person and a state (Keating, Cairney and Hepburn 2009, 54). Territory-based links imagine citizens connected within the borders of the state, which is deemed to create associative bonds that qualifiy individuals to be members of that polity. This type of link has its origins in feudalism where birth in a territory meant allegiance to a monarch. This allegiance was subsequently engrained in the citizenship rules of the British Common Law system, which spread to North and South America through colonial conquest. Its most manifest contemporary rendition is unconditional automatic citizenship by birth in the country's territory, irrespective of the parents' nationality (ius soli). Citizenship acquisition by ius soli therefore ties birth in the territorial spheres of the state to membership.
Yet the recognised connections created through territory go beyond ius soli. In terms of governance, these territory-based links are ingrained in the notion of residency, considered the 'functional grounds' of membership (Shachar 2009). Territorial connections can be based on the residency of the parent, or of the individual seeking to become a citizen. Through ordinary naturalisation and admission of long-term residents, residency acts as a proxy for social and civic bonds. Even for those granted admission under international law (refugees, asylum seekers and stateless persons), the pathway to citizenship also falls under territorially based links as multiannual residency is the primary acquisition requirement. Residency is defined differently under national jurisdictions, and the creation and maintenance of connections to the state is commonly accompanied by a further qualification of residence as effective, habitual, or permanent. As a result of such qualifications, territory-based links can be lost after a certain period of residency abroadthe assumption being that the connection to the state is broken after physical presence is severed.
Third, recognised links can be conceived through ties to a state's citizen, or someone seeking naturalisationwhat we call 'sponsorship'for example, through facilitated naturalisation requirements for spouses or extension of naturalisation to family members (e.g. reduction of residency periods, waiving of application fees and simplification of bureaucracy). The rationale, in this case, is that a person's connection to the state can be (partly) extended or transferred to spouses, partners, adopted children, or immediate family members. Such links are believed to facilitate the physical and cultural reproduction of the boundaries of the nation-state (Kristol and Dahinden 2019) and are tied to the idea that the family is the essential building block of society. Sponsor-based ties depend a great deal on the definition of a 'family'. Throughout the twentieth century, spousal transfer and acquisition for adopted children have been restrictive as it is largely practised from a heteronormative perspective. For example, until after the First World War, a married woman's citizenship was almost universally dependent on her husband in the Western hemisphere (Volpp 2017). While having evolved along with the international human rights standards of the twentieth century (Bonjour and Block 2015), spousal transfer has spurred many contestations as societal notions of 'the family' change. This conception of link defines what the state views as 'acceptable' and 'legitimate' relationships.
Finally, merit can be constitutive of a 'link'. There are different views of what exactly is considered as 'merit'. 1 However, it is connected to an idea of deservingness based on an individual's exceptional abilities and talent (Joppke 2021;Sandal 2020;Shachar 2011). While it is broadly considered that 'positions and rewards should be distributed according to individual merit' (Young 1990, 200), the exact criteria for measuring merit for citizenship acquisition and loss has changed throughout history. The concept is historically rooted in the Ancient Greek model of citizenship, which saw membership as contingent on an individual's responsibilities toward the state (Riesenberg 1992). Therefore, membership was a way to guarantee loyalty that could be lost when certain expectations were not metsuch as, in cases of fraud or treachery. Similar models of membership can be traced back through the Middle Ages and to the rise of nations in the seventeenth and eighteenth centuries (Prak 2018). Nowadays, we see the legacy of such membership conceptions in rules for granting citizenship to those deemed to serve or advance national intereststhrough special achievements in academia, sports, science, technology, arts (Shachar 2011;Joppke 2018), through public services, or financial assets (Dzankic 2019;Surak 2021). The underpinning idea is that outstanding accomplishments (as defined by the state) are a recognisable establishment of a connection. This 'meritdriven' approach is also applied in citizenship revocation. In these cases, a link is severed if a person is involved in actions considered harmful to the state's interests such as, service in a foreign military or in an act of attrition (Joppke 2021).
Our typology thus points to how a state's citizenship laws can follow similar logics in the realms of citizenship acquisition and loss for each kind of state-citizen link. These links then come together to form an external boundarywhere the shell of the state can be hard in some places and soft in others depending on the type of link a person has to it. Such a framework can inform our understanding of membership in multilevel states and supranational organisations far better than previous typologies on citizenship. 2 States have unique political, social and economic interests that can play out in certain domains of their citizenship legislationwhile other domains remain relatively static. These interests continue to coexist in a multilevel context.

The domains of citizenship links in the Member States of the EU
While the status of EU citizenship confers equal rights onto the beholdershow that status is achieved and when it is taken away differs based on the individual Member State (and their own membership incentives) that guards the gateway. Thus, understanding the substantive dimension of EU citizenship requires an awareness of not only the fact that this status is created by the nationality laws of the 27 Member States, but also of the types of links that these laws articulate. In the following sections, our typology explains the coming together of the citizenship legislation of the EU Member States in constituting the cragged boundaries of EU citizenship.

Lineage-based links
Automatic citizenship by birth to parents who are citizens within the country's borders (ius sanguinis) is one of the most common mechanisms for the attribution of citizenship worldwide. It exists in all EU Member States. In addition to this, some Member States also extend the scope of ius sanguinis beyond national territories, to children born abroad to a citizen parent, or even to future generations. They delimit the extraterritorial scope of EU citizenship by creating populations who have been born and reside outside EU territory but who enjoy the rights attached to Union citizenship rights, including free movement inside the EU, subsidiary diplomatic and consular protection and, in some cases, voting rights in European Parliament elections (European Commission 2019). The uneven boundaries of EU citizenship in this domain thus result from two factors where there is strong variation between Member States: (1) the extent to which ius sanguinis is applied to generations born abroad; and (2) the additional conditions (or restrictions) that Member States require to access citizenship.
Belgium, Croatia, Cyprus, Estonia, Finland, Germany, Ireland, Malta, Portugal and Slovenia have limitations to the automatic conferral of citizenship through descent to children born outside of their territory. These limitations include mandatory registration with a local consulate in the country of destination or a foreign birth register in the country of origin (Belgium, Cyprus, Germany, Ireland, Portugal), wedlock restrictions (Malta and Finland), or the requirement that both parents are citizens (Croatia, Estonia, Slovenia). Limitations to transmission of citizenship iure sanguinis to further generations reflect a concern that citizenship should be based on a different form of a link to a country. From the perspective of some states, this link cannot be presumed for children born abroad whose grandparents (or more distant ancestors) have emigrated. By contrast, Greece, Italy and Ireland follow a 'familistic' model where intergenerational transfer is unlimited (Zincone and Basili 2010). The logic is that bloodbased membership cannot be brokeneven when the population lives for generations outside the state. In old emigration societies such as Italy, these provisions strengthen cultural and economic ties with the diaspora (Finotelli, La Barbera and Echeverría 2017). For example, between 1998 and 2012, Italy granted citizenship to 1 million descendants born abroad. Thus, the conditions and limits to ius sanguinis acquisition exacerbate the uneven boundary of EU citizenshipwhere those born with the ancestry from certain Member States will pass down EU citizenship to generations of children born abroad and those born with another will not have that same privilege.
Beyond descent as lineage, belief in a common origin is considered as functional grounds for citizenship acquisition in most post-communist Member States. Bulgaria, Croatia, Hungary, Lithuania, Poland, Romania and Slovakia all have provisions in their citizenship legislation facilitating the admission of new citizens based on presumed 'origin', descent, or associations (GLOBALCIT 2022). Szabolcs Pogonyi has shown how 'ethnonationalist political projects that emphasise the 'right of blood' may lead to the naturalisation of individuals who have very weak cultural or linguistic ties to their putative 'homeland'' (2022, 1). Pogonyi finds that despite international and EU norms of nondiscrimination, the current legal framework in the EU does not exclude the possibility of selecting citizens based on ethnic proximity. This, however, could be a future site of contestation, as more generations are born abroad with EU citizenship and weak social ties to the Member State.
In contrast to the many lineage-based connections for acquiring citizenship, there exist only two modalities in which this type of a link can be lost. First, the lineagebased connection is lost if descent is annulled. Finland, Germany, Luxembourg and the Netherlands may remove citizenship of a child (in Finland and Germany before the age of five), respecting the relevant statelessness safeguards. Second, lineage-based links are broken in cases where the person does not renounce a foreign citizenship after receiving citizenship iure sanguinis. This is the case in Germany, Latvia and Lithuania, with exceptions for 'permitted countries' where dual citizenship is tolerated when both citizenships are acquired at birth. Such 'permitted countries' are politically, historically, economically, or culturally related. Hence, in these cases, the boundary of lineagebased links is internally uneven where some groups of citizens can access dual citizenship while others cannot.
One of the reasons that there is such diversity among the Member States in the lineage realm of links is that the connections which represent the intergenerational continuity, real or presumed consanguinity, or cultural ties have been relatively 'off limits' as it represents 'who one is' in relation to the Member State (see Pasetti 2021 for example of the Italian case). Notwithstanding, one area where lineage-based links have already started to change are related to the conferral of birthright citizenship to children born to parents whose interpersonal relationships reflect non-heteronormative gender or sexual identities. This can be seen indirectly as a result of the application of community law and international human rights standards in recent judgments of the CJEU, which otherwise has limited powers in the domain of citizenship transmission. 3 The case concerned the failure of Bulgaria to recognise citizenship to children born to same-sex couples, due to national restrictions as to who the parents of a child may be (only a man and a woman) (Cabral 2021; de Groot 2021). 4 Due to its limited power, the CJEU could not rule on citizenship directly, but the effect of the ruling has consequences for citizenship attribution for children of same-sex couples who decide to exercise their rights under the treaty, thus protecting lineage-based citizenship acquisition for those who establish it within another jurisdiction.
All of these developments indicate that in the realm of lineage-based links states continue to follow a national models' approach with limits on the intervention capacity on citizenship laws from international and EU processes. The boundary of EU citizenship in this realm of link is likely to remain uneven. Even so, it could be a site of future contestation in cases where (1) LGBTQ+ couples are not able to transmit citizenship with the same facility as heterosexual couples; (2) states do not limit ius sanguinis to citizens born abroad; and (3) where states offer citizenship for those who share a 'common origin' yet have tenuous social ties to the Member State.

Territory-based links
There are uneven pathways to EU citizenship in the context of some of the rules falling in the domain of territory-based connections, particularly as regards citizenship by birth in the territory of a Member State (ius soli). Unconditional ius soli is uncommon as the principal mode of citizenship attribution in the EU and we find no states who apply a purely automatic ius soli except in cases of foundlings or otherwise stateless children. Germany, Greece, Ireland and Portugal offer citizenship at birth to children of residents (regardless of their parents' nationality but accompany it with further conditions such as legality of status, years of residence, or schooling) (Erdilmen and Honohan 2020). Equally, Belgium, France, Greece, Luxembourg, the Netherlands, Portugal and Spain grant citizenship ex lege to children born in a country to a parent who was also born in that country. Only France practices a 'double ius soli' principle where the children of a foreign national parent born in France will receive citizenship automatically without additional qualifications like in other Member States which do not presume a link based on birth alone. In France, this provision was historically used to expand those that were eligible for the draft. The contemporary context, however, gathers that sufficient links are established for automatic acquisition by the birth of two generations born in the territory.
All the Member States offer residence-based naturalisationhowever, the corresponding conditions and the different meanings of residency under national legislation create uneven pathways. In terms of time, Member States range from requiring between three and eight years of residence (Dzankic, Mantha-Hollands, and Baubock 2020). Further conditions tied to this status may entail continuous residence (in Spain, 5 out of 10 years must be continuous, i.e. the applicant must have not had periods of absence from the country's territory longer than 6 months) or permanent residency (in Austria, applicants need to be permanent residents for five years; the permanent residence permit is commonly obtained after five years of continuous residence). Such residency conditions for naturalisation are typically combined with civic integration tests, language requirements and employment contracts (Vink and . While there is no EU wide harmonisation regime for integration requirements, there are some soft-law measures (de Waal 2021, 20). In 2004, the Council of the European Union adopted a non-binding decision on the Common Basic Principles for Immigrant Integration Policy in the EU that emphasised language, history, and mutual participation as key principles of successful integration. These additional conditions in some cases have begun to reshape the substance of what is understood as residency for the purpose of naturalisation across the EU. In countries such as Denmark, France and Italy, residency is linked to employment or independence from social welfare; hence territory has become intertwined with economic self-sufficiency. In other words, while laws may stipulate the condition of legal residence (physical presence within the boundary of the state in most cases), additional factors related to integration are becoming increasingly important when assessing territory-based links to the state.
One area of citizenship law where we find the reverse derivative quality (where EU citizenship is the gateway to Member State citizenship) is the facilitation of naturalisation by means of waiving conditions that commonly attest to links (mandatory residence periods, language knowledge, civic test, absence of conviction, oath, ceremony and renunciation of another citizenship) for nationals of all other Member States. As of 2022, this happens in Austria, Bulgaria, Czech Republic, Germany, Greece, Italy, Latvia and Romania, which facilitate the acquisition of citizenship for nationals of another EU Member State by reducing residence periods or waiving the renunciation requirement (GLOBALCIT 2022). 5 While ordinary naturalisation applicants need to meet the 10-year residence condition in Austria, this requirement is reduced to six years for EU citizens. Italy and Romania apply a four-year residence condition to nationals of other EU Member States, while requiring 10 and 8 years of residence from third-country nationals. The Czech Republic and Greece both decrease the mandatory residence period from seven to three years. Bulgaria, Latvia and Germany apply the same residence and other conditions to EU nationals but waive the requirement to renounce a previously held EU citizenship. In all these cases, the Member States have voluntarily introduced preferential provisions for nationals of other EU countries. The extension of citizenship is thus conceived on the basis of the territory associated with the EU rather than on any specific national membership, or other kinds of historical or cultural links. This is further supported by the fact that facilitated acquisition for other EU nationals is independent of the length of membership, size of the country, or its geographical location.
In a similar vein, the way that territory links a person to a state, prolonged residency abroad can sever that link from the state's perspective. European citizenship offers freedom of movement and settlement within the territorial boundaries of the EU. Therefore, we observe similar rules in Member States that would normally revoke citizenship in the case of prolonged residency, but do not do so if the person lives within the EU. For example, this is the case in the Netherlands, where individuals who have resided outside the EU for a continuous period of 10 years will lose their Dutch citizenship; instead, residence in another EU country counts as residence in the Netherlands (Van Oers et al. 2010, 17). A further norm is the requirement for application of tests for balancing interests and proportionality, resulting from recent CJEU jurisprudence. 6 Yet this norm is incomplete as the implementation of these tests differ among the EU countries where citizenship can be lost after prolonged residency abroad (Belgium, Cyprus, Denmark, Finland, Ireland, Malta, Netherlands, Spain and Sweden). The scope, procedure and requirements of revocation laws vary significantly. In terms of scope, Cyprus, Ireland and Malta apply this ground for withdrawal only to naturalised citizens who reside abroad for more than seven years. In Belgium, Denmark, Spain and Sweden, only citizens born abroad may lose their status if they do not reside in the respective Member State. For the procedure, in Belgium, Denmark, Finland, Netherlands, Spain, Sweden loss of citizenship occurs without the explicit consent of the person, i.e. through an ex lege loss mandated by a public authority; whereas, in Cyprus, Ireland and Malta procedural safeguards in place are in conformity with EU law and loss is subject to an individual decision by authorities (Handoll 2010;Buttigieg and DeBono 2015;Trimikliniotis 2015).
Such dynamics between the evolving legal provisions and the mechanisms for their implementation indicate how citizenship laws in the domain of territory-based link are slowly converging to even out the boundary of EU citizenship in two respects: (1) in citizenship acquisition rules for citizens of other Member States; and (2) in the application of the principle of proportionality for citizenship loss based on discontinued residence. All Member States offer the possibility for residence-based naturalisation which is accompanied by conditions that are similar in scope (e.g. residence, language, civics) but different in substance (e.g. duration, level, score). In this respect, the similarity in conditions perhaps reflects the 'soft law' measures of the EU; however, the boundary of EU citizenship remains uneven as the degree to which these conditions are applied in practice reflect 27 conceptions of what is entailed to become a citizen.

Sponsorship-based links
Personal connections to a state's citizen may entail a waiver of some of the conditions for naturalisation that would otherwise be applicable to ordinary, residence-based citizenship acquisition. In cases of naturalisation through marriage, partnership or adoption, the conception of links rests on the presumption that a personal relationship will result in socio-cultural integration. In practice, all EU Member States except for Estonia, Latvia and Poland facilitate citizenship acquisition for spouses and partners of citizens. The boundary of EU citizenship is shaped in this domain of citizenship legislation based on three factors: (1) the number of years required for residency before citizenship based on marriage or adoption is granted; (2) whether other naturalisation requirements are waived; and (3) if provisions extend to LGBTQ+ partnerships. For instance, Austria, Belgium, Bulgaria, Cyprus, Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Malta, Netherlands, Portugal, Romania, Slovakia, Slovenia, Spain and Sweden all reduce the residence requirement for spouses and adoptees compared to ordinary naturalisation. Spain additionally waives the obligation for spouses to renounce their other citizenship. This indicates some convergence in conceiving of linkages through personal connections.
Even so, there are major differences in conceiving of sponsor-based connections as regards the recognition of LGBTQ+ marriage (Higdon 2019). All of the Member States in North and West Europe have either legalised same-sex marriages or introduced other legal mechanismssuch as civil unionsto equalise the rights of individuals in partnerships reflecting all gender and sexual identities. This recognition in Eastern European states is yet to fully exist. In 2017, a referendum failed in Slovenia to abolish the traditional definition of marriage enshrined in the country's constitution. Subsequently, the country adopted a separate Act Concerning Partnership. Equalising the rights of LGBTQ+ unions to those applicable to 'marriage' in most legal domains, the Act allowed foreign partners of Slovenian nationals to benefit from facilitated naturalisation. Yet similar legislative changes did not take place across the board in Eastern European countries: Bulgaria, Latvia, Poland, Romania and Slovakia do not acknowledge the rights of same-sex partners. However, following the 2018 Coman (Case C-673/16) judgment of the CJEU, Member States are bound to grant residence rights to same-sex partners of EU nationals regardless of whether these partnerships are recognised in the respective Member State or not. While such partners will be unable to benefit from the facilitated spousal transfer (except in Latvia and Poland), they would qualify for residence-based naturalisation on the grounds of entry and stay rights received through partnership. This development emanates from an increasing recognition of the fact that a spectrum of interpersonal relations constitutes links with the state as much as the traditional notion of 'marriage' does. Thus, while CJEU jurisprudence indirectly attempts to even out the pathway to citizenship for spouses of LGBTQ+ partners, this realm of link will remain uneven until all Member States provide equal pathways for partners of all types of relationships.
In the realm of citizenship loss in this domain of link, we find cases where children who have naturalised through their parents can lose their citizenship if their parent loses citizenship. This is the case in Austria, Belgium, Croatia, Denmark, Finland, Germany, Greece, Italy, Lithuania, Luxembourg, Netherlands, Poland, Romania, Slovakia and Slovenia. In Austria, Bulgaria, Finland and Germany this can only occur if it is proven that the citizenship of the parent was acquired through fraud. In Belgium, Bulgaria, Croatia, Luxembourg, Poland, Romania and Slovakia a child can lose citizenship if the parent renounces said citizenship. However, each country provides an age restriction or requires the child's consent (Dzankic, Mantha-Hollands, and Baubock 2020). For this link, the potential for loss, the differing age requirements, and scope of procedure all perpetuate the uneven boundary of EU citizenship. While this domain of link will likely continue to remain uneven as it represents individual identities and personal relationships in relation to the Member State, it may be a site for future contestation if it is found that the conditions for the loss of citizenship do not align with the principle of proportionality or with principles of non-discrimination towards non-heterosexual couples.

Merit-based links
Among EU Member States three types of naturalisation provisions reflect merit-based links between individuals and states: special achievements, public service and financial assets. In such merit-based citizenship acquisition, states recognise the past or the potential for a future exceptional contribution and facilitate admission by waiving one or more naturalisation conditions that would be applied to ordinary applicants. 7 Admission based on national interest, special achievement, or public service is a widespread provision. The substance of regulation reflects an assemblage of different meanings of 'national interest' or 'achievement'; thus, these provisions create uneven grounds for which EU citizenship is granted. Most Member States have provisions for naturalisation through special achievements (with the exception of Denmark, Finland, Poland, Spain and Sweden). France, Germany, Greece, Ireland and Italy facilitate citizenship acquisition for individuals who performed military or civil service for the state. Legislative provisions in Austria, Bulgaria, Slovakia and Slovenia that specifically refer to 'commercial' or economic 'interests' may also be used for naturalisation based on investment, with a substantial or nearly full waiver of other conditions. 8 These naturalisation provisions allow a significant amount of discretion and what exactly constitutes the array of 'interests' or 'services' has varied over time.
Between 2013 and 2022at different points in time and in different formsthree EU Member States (Bulgaria, Cyprus and Malta) had specific provisions that enabled naturalisation on the basis of investment, derived from the possibility to naturalise on the basis of merit (Dzankic 2019;Parker 2017). Investor citizenship programmes in Bulgaria and Cyprus have been terminated on 24 March 2022 and 1 November 2020; at present, Malta is the only EU country running such a programme. These 'golden passport' schemes have been much criticised by different EU institutions and other Member States, primarily, due to the security risks that they create (Scherrer and Thirion 2019); issues that have become increasingly salient in the context of the conflict in Ukraine (European Commission Press Release, 28 March 2022). Yet beyond the governance-related issues, the European Commission has based its ongoing infringement proceedings against Malta on the argument that 'the granting of EU citizenship in return for pre-determined payments or investments, without any genuine link to the Member State concerned, is in breach of EU law' (INFR 2020). The Union has limited competence in the domain of citizenship, and the notion of 'genuine link' is much contested as a legal principle and unclear as the normative ground for citizenship attribution (Van den Brink 2018). However, the Commission's approach implies that while the terrain of EU citizenship may be uneven, certain practices can still be contested.
While each Member State has differing conceptions of exceptional merit from the point of view of citizenship acquisition, we find similarities in the procedures for the modes of citizenship lossthe revocation of citizenship due to fraud, disloyalty and service in a foreign army. This is largely attributable to compliance with CJEU jurisprudence. The Rottman judgment (C-135/80) established that in the case of fraud resulting in the loss of national citizenship (and as a consequence EU citizenship), states would need to apply a test of proportionality. However, the application of this test is done to varying degrees. Austria, Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Portugal, Romania, Slovakia, Slovenia and Spain have provisions for withdrawal of citizenship or nullification of naturalisation given fraudulent acquisition (GLOBALCIT 2022). Elements of the principle of proportionality are built into the revocation procedure, through four types of provisions. The first are time limits to when the state can investigate fraud cases. Belgian law, for example, stipulates that individuals may have a 'reasonable period of time' to recover their original nationality. The second are safeguards that citizenship can be revoked only when the fraud has had a 'conclusive' effect on naturalisation, which exist in Cyprus, Denmark and Finland (Lepoutre 2020). Third, is the protection of citizenship status for those associated with the offender, such as spouses or children. For example, Spanish law ensures that any uninvolved third parties will be protected from harmful effects caused by the revocation decision. And fourth, most apparently, is the assessment of the individual's overall circumstances. This is the case in Finnish law, which stipulates that a decision to revoke citizenship must be 'based on an overall consideration of the person's situation' and 'the culpability of the act and the circumstances in which it is committed, and of the ties with Finland of the person who has made the application or declaration' (Lepoutre 2020).
In the contemporary landscape, loss procedures for 'disloyalty' and 'service to a foreign military' are primarily used in the context of terrorism-related offences although recently, has also applied to foreign fighters in Ukraine and Russian/Belarusian nationals (Joppke 2016;Fargues and Winter 2019). 9 Since 2016, Belgium, Denmark, France, Germany, Italy and the Netherlands have expanded legislation in this regard. The application of this legislation differs depending on the Member State. In Denmark the requirement is vague stating that the person must have 'acted in a manner that is seriously prejudicial to the vital interests of the country' whereas, in Germany it is more specific, highlighting that the individual must be actively involved in a 'terrorist militia'. In some Member States the provision will only apply to naturalised citizens (Dzankic, Mantha-Hollands, and Baubock 2020). We find that some Member States have also structured the implementation of revocation procedures around the principle of proportionality. For example, in France, there is a time limit in which such a crime can result in the loss of citizenship (1 year before to 10 years after acquisition of French citizenship). In Italy and Latvia, citizenship revocation must follow a judicial decision thus assuming that measures of proportionality will be tested in the criminal proceedings. In this domain of loss, we, therefore, find that while there are disparate mechanisms for constructing membership, the procedures implemented in order to comply with CJEU jurisprudence work to even out the EU citizenship boundary.

Conclusion
Member States continue to be the gatekeeper of citizenship acquisition and loss in the EU and thus chart an uneven terrain of EU citizenship. While the political and mobility rights conferred by the treaties related to Union citizenship remain the same for all EU citizens, there still is a disparity as to who is (un)recognised as a citizen of the Union. This can be observed, albeit to different degrees, in the four conceptions of state-citizen links, established through (i) lineage; (ii) territory; (iii) sponsorship; and, (iv) merit.
First, in the case of lineage-based links we see the persistence of national models for ensuring the inter-generational continuity of citizenship. However, the unevenness in this domain comes from the number of generations citizenship is extended to for descendants born abroad, presumed lineage, and in cases related to the automatic transfer of citizenship for children of LGBTQ+ parents. Second, with territory-based links we find that all Member States provide for residence-based naturalisation however, this is done with varying time periods and conditions. Thus, the pathway to EU citizenship depends on which Member State an applicant is a resident. Similarly, the different definitions of legal residence that are taken into account for the application of ius soli shape the uneven contours of EU citizenship. However, we observe a reverse effect for the 'extension' of ius soli after birth through facilitated naturalisation requirements for other EU nationals. This is not surprising, as certain fundamentals of the EU (freedom of movement and the single market) are based on conceiving the EU as a bounded territory. The concept of membership in the EU is thus, derived from its conception of territory as 'the place of a collective' (Shuibhne 2019, 268). Moreover, we find in terms of citizenship loss, the application of the principle of proportionality helps to even out the EU's membership boundary. Third, we observe two main variations in the realm of sponsor-based links, to the extent the transfer is extended (facilitation of other conditions) and whether LGBTQ+ partnerships are included for spousal transfer. Fourth, with merit-based links, in the domain of citizenship acquisition we observe that similar legislative provisions articulate different national interests and priorities. In the domain of loss, however, procedural safeguards are becoming increasingly coherent across the Member States, as the loss of national citizenship also results in the loss of EU citizenship.
The 'uneven' substance of EU citizenship is the result of a combination of factors affecting the regulation of membership across the different types of links. These factors include national interests and preferences, as well as broader European and international processes. They manifest themselves differently on the types of citizenship links. For instance, in the case of lineage-based connections, legislation is largely attributable to the national conceptions of membership (Howard 2009;Brubaker 1992). This is the case because, traditionally, citizenship connections based on 'who one is'by blood or personal connectionare more resilient as they are conceived to ensure intergenerational continuity of the national community (Vink and Bauböck 2013) or preserve elements of its historical role. Even so, our research shows that the rulings of the CJEU may affect consanguineous or presumed heritage-based citizenship transmission in view of the evolving individual identities, the ensuing relationships and their outcomes. A similar effect is expected in cases of sponsor-based links, where changes in legislation have been driven by the evolution of international and European human rights norms.
In those domains of citizenship law where links between individuals and states are conceived through territory and merit, links are created and lost through place and action. It is within these types of links that we find increasing norms among the Member States. The domains of citizenship law that reflect the links constituted through territory or merit are more susceptible to change because they reflect the states' response to emerging challenges, in the form of reconstructing its territorial scope, reframing the relationship between population or territory, and thus redefining 'what one does' for the state.
Member States remain distinct national membership communities but have similarities and differences within each of the types of linkages that make up their individual and collective frontiers. Unlike the existing scholarship, which has taken a top-down approach and looked at citizenship in the EU's Member States through the lenses of EU citizenship, we take a bottom-up method. That is, we explore how EU citizenship is conceived through the substantively different state-citizen connections encapsulated in Member States' citizenship laws. Our lens, which has shown an uneven construction of EU citizenship, provides an alternative angle in which to study how multi-level citizenship is constituted by the coming together of distinct, and often contradictory, aspects of national membership policies. Notes 1. While some scholars associate 'merit' with conditions for naturalisation i.e. ex-post residency within the state (Ganty 2021; Ammann 2020); our approach considers 'merit' as irrespective of residency. The state decides whether an achievement, ability, or talent is worthy of granting citizenship which is not based on any predefined requirements but vague discretionary standards of 'national interest'. 2. Earlier citizenship typologies have focused on other research agendas, such as levels of inclusion and exclusion (Schmid 2020;Howard 2009;Janoski 2012;Koopmans, Michalowski and Waibel 2012), the functional purposes of citizenship laws (Vink and Bauböck 2013), claims-making (Bloemraad 2018), and national models (Brubaker 2009;Soysal 1994). Much of this work has been constrained to citizenship acquisition for immigrants and abstained from highlighting variations within states' citizenship regimes comprehensively (Waldrauch 2006 States or citizens of their formal colonial empires. For example, in Spain where European citizens must wait 10 years for naturalisation, Portuguese citizens can reduce the residency requirement by 2 years. This is also the case for citizens of Latin American countries, Andorra, Philippines, or Equatorial Guinea. In Portugal, citizens of Angola, Brazil, Cape Verde, Guinea Bissau, East Timor, Macau, Mozambique, and São Tomé and Príncipe can benefit from a facilitated pathway to citizenship. The same is true among the Nordic countries. 6. See most recent CJEU decision regarding the loss of citizenship due to prolonged residency abroad in Tjebbes C-221/17. 7. Most EU Member states fully or substantively waive the residence condition, as well as subsistence criteria, or any language or civics. Conditions related to good character are commonly retained. 8. Bulgaria, Slovakia, and Slovenia may waive all other material conditions for the acquisition of citizenship based on 'economic' interest (except good character). In Austria, under the constitutional provision 10(6), all material conditions could be waived for major investment (no guidelines as to the amount or type). Alternatively, under article 11(4)(4), a reduction from 10 to 6 years of residence is foreseen for economic contribution, but all other naturalisation conditions are retained. 9. In the context of conflict in Ukraine, passports obtained through investor citizenship programmes have come under scrutiny. In its recent recommendation, the European Commission (2022) has called for an assessment of whether passports of Russian and Belarussian nationals subject to sanctions and substantively supporting Russian aggression in Ukraine should be withdrawn. Moreover, Austria announced that it would revoke citizenship for any of their nationals who entered military service in Ukraine, Germany stated that it would 'examine a solution' for its dual Ukrainian-German citizens, and Latvia amended the law that deprives citizenship for joining the armed forces of a foreign state for Ukrainian combatants (Van der Baaren et al. 2022, 11).